interlanguage pragmatics bardovi-harlig&hartford lb

INTERLANGUAGE PRAGMATICS

Exploring Institutional Talk

Second Language Acquisition ResearchTheoretical and Methodological Issues

Susan M. Gass and Jacquelyn Schachter, Editors

VanPatten

Tarone/Gass/Cohen

Schachter/Gass

Birdsong

Ohta

Major

• Processing Instruction: Theory,Research, and Commentary

• Research Methodology in SecondLanguage Acquisition

• Second Language Classroom Research:Issues and Opportunities

• Second Language acquisition andthe Critical Period Hypotheses

• Second Language Acquisition Processesin the Classroom: Learning Japanese

• Foreign Accent: Ontogeny andPhylogeny of Second LanguagePhonology

MONOGRAPHS ON RESEARCH METHODOLOGY

Gass/Mackey

Yule

Markee

Dornyei

OF

Gass/Sorace/Selinker

Gass/Selinker

• Stimulation Recall Methodology inSecond Language Research

• Referential Communication Tasks

• Conversation Analysis

• Questionnaires in SecondLanguage Research: Construction,Administration, and Processing

RELATED INTEREST

• Second Language Learning DataAnalysis, Second Edition

• Second Language Acquisition: AnIntroductory Course, Second Edition

INTERLANGUAGE PRAGMATICS

Exploring Institutional Talk

Edited by

KATHLEEN BARDOVI-HARLIGBEVERLY S. HARTFORD

Indiana University

LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS2005 Mahwah, New Jersey London

The camera copy for the text of this book was supplied by the editors.

Copyright © 2005 by Lawrence Erlbaum Associates, Inc.All rights reserved. No part of this book may be reproduced inany form, by photostat, microform, retrieval system, or any othermeans, without the prior written permission of the publisher.

Lawrence Erlbaum Associates, Inc., Publishers10 Industrial AvenueMahwah, New Jersey 07430www.erlbaum.com

Cover design by Kathryn Houghtaling Lacey

Library of Congress Cataloging-in-Publication Data

Interlanguage pragmatics: exploring institutional talk / edited by Kathleen Bardovi-Harlig,Beverly S. Hartford.

p. cm.Includes bibliographical references and index.ISBN 0-8058-4890-8 (cloth: alk. paper) — 0-8058-4891-6 (pbk. : alk. paper)

1. Discourse analysis—Social aspects. 2. Conversational analysis. 3. Pragmatics.4. Interlanguage (Language learning) I. Bardovi-Harlig, Kathleen, 1955- II. Hartford, Beverly S.

P302.84.1584 2004306.44—dc22

2004053192

Books published by Lawrence Erlbaum Associates are printed on acid-free paper,and their bindings are chosen for strength and durability.

Printed in the United States of America1 0 9 8 7 6 5 4 3 2 1

www.erlbaum.com

Contents

1

2

3

4

5

6

7

8

IntroductionKathleen Bardovi-Harlig and Beverly S. Hartford

Institutional Discourse and Interlanguage Pragmatics ResearchKathleen Bardovi-Harlig and Beverly S. Hartford

Writing Center Interaction: Institutional Discourse and the Role ofPeer Tutors

Jessica Williams

Negotiating an Institutional Identity: Individual Differences in NSand NNS Teacher Directives

Lynda Yates

Before, During, and After the Event: Getting the Job (or Not)in an Employment Interview

Julie Kerekes

Discourse Strategies in the Context of Crossculrural InstitutionalTalk: Uncovering Interlanguage Pragmatics in the UniversityClassroom

Catherine Evans Davies and Andrea E. Tyler

English for Specific Purposes and Interlanguage PragmaticsElaine Tarone

Using Moves in the Opening Sequence to Identify Callersin Institutional Settings

Tara Leigh Gibbs

Practical ConsiderationsKathleen Bardovi-Harlig and Beverly S. Hartford

Author Index

Subject Index

1

7

37

67

99

133

157

175

201

223

227

V

This page intentionally left blank

Introduction

Kathleen Bardovi-HarligBeverly S. Hartford

Indiana University

Discussions of research design in interlanguage pragmatics reveal a tensionbetween the desire for highly controlled production tasks that yield comparablelanguage samples and the desire to integrate the investigation of authenticdiscourse into studies of interlanguage pragmatics. In spite of the interest thefield has expressed in authentic discourse, controlled elicitation tasks stilldominate data collection. This book introduces interlanguage pragmaticsresearchers to institutional talk, talk that occurs in the course of carrying out aninstitution's business, usually between an institutional representative and aclient. The collection and analysis of institutional talk—one form of authentic,consequential discourse—meets the field's methodological requirements ofcomparability, predictable occurrence of pragmatic features, high rates ofoccurrence, and relative ease of data collection. Other advantages of exploringinstitutional talk include the availability of participants for retrospective inter-views and of consultants for interpretation. The investigation of institutional talkcan be used to study the influence of context, the interaction of pragmaticfeatures, the effect of timing and escalation, the comparison of goals andoutcomes, and acquisition of linguistic and pragmatic features. Institutionalsettings also afford researchers the opportunity to observe the acquisition ofinstitutional rules themselves, which represent a microcosm of culture.

In order to explore these advantages fully, we have assembled previouslyunpublished studies in this volume that address both interlanguage pragmaticsand institutional talk. Taken together, these studies demonstrate the benefits ofstudying—the context of authentic and consequential talk—the traditionalvariables of status, directness, social distance, imposition, and others typicallyinvestigated in interlanguage pragmatics. Many chapters in this volumeintroduce new concepts, thus expanding the range of interlanguage pragmaticsresearch beyond what can be investigated in production questionnaires.Additional concepts that are investigated include trust, individual variationacross multiple turns and interactions, authority and equality (in balance todominance and solidarity), and discourse style. The chapters also challengereaders to reconsider the primacy of the dichotomy between native speaker andnonnative speaker. Importantly, these chapters demonstrate that success (andlack of success) is found on both sides of this linguistic divide. We are hopefulthat these fine-grained analyses will contribute to a better integration of thenonnative speaker in pragmatics research. The potential for extended analysis ofboth native and nonnative pragmatics in authentic interaction helps placeemphasis on individual pragmatic strategies that are (un)successful which maybe more or less prevalent in one speaker group than another. This will lead to agreater sophistication in our description of why native speakers may be more

1

2 Bardovi-Harlig and Hartford

successful, but avoids the pitfall of attributing "pragmatic competence" to nativespeakers (NS) and "lack of pragmatic competence" to normative speakers(NNS).

To this end, the studies reported in this volume are set in a variety ofinstitutions in which both NS and NNS participate as institutional represen-tatives and clients. The range of settings investigated in this volume include auniversity writing center, a university physics laboratory, a variety of secondaryschool classrooms, an employment agency, and a four-star conference hotel. Thedifferent settings show the importance of locating research in talk that native andnormative speakers actually perform. Speakers engage in talk in settings that areimportant to them and studies of situated language observe people engaged intalk in institutions which they frequent. This then eliminates the methodologicalquestion of constructing relevant scenarios in elicitation tasks and provides theopportunity to observe actual outcomes.

Cataloguing differences between NS and NNS is ultimately insufficient: Wemust also understand which differences are important (Kasper, 1997). One wayto do this is to study outcomes. The study of consequential talk found ininstitutions furthers that goal: Native and normative differences can beunderstood in light of what pragmatic strategies seem to contribute to and whichseem to impede the interlocutor's success at the institution. Institutional eventshave outcomes. An advising session may result in a signed registration form, awriting tutorial may result in a better paper, an interview at an employmentagency may result in a job referral—or not. Outcomes viewed as endpoints arenot the only measure of success, however. There may be characteristics of theencounter itself that indicate a felicitous or infelicitous interaction: how long ittakes to negotiate, what follow-up questions are asked during the encounter, andspeaker dominance, for example.

Whereas a predetermined research design may set up a dichotomy betweennative and normative speakers, we may find by observing authentic interactionsthat such a dichotomy is only one variable among many. The NS-NNS speakerdivision may be less important than shared interests or common background,education, job types, gender, as studied by Kerekes, individual styles,investigated by Yates, or expertise, as discussed by Tarone and Gibbs. Some ofthese differences in viewing the participants are evident in the descriptors usedby the various authors, including the familiar NS-NNS designation, as well asNE (Native English) and L2 Writers used by Williams, and novice and expertmembers of a discourse community, employed by Tarone and Gibbs.Participants may also be referred to by their roles, such a student and teachingassistant, as in the chapter by Davies and Tyler.

This volume is organized into eight chapters. The first chapter provides anoverview of research that has been done in interlanguage pragmatics in thecontext of institutional talk. This is followed by six chapters that report onstudies conducted at institutions and that illustrate a number of advantages tousing institutional talk as a source of information informing the acquisition of

Introduction 3

L2 pragmatics. The volume ends with a chapter that addresses the practicalconsiderations of conducting research on pragmatics at institutions.

In the first chapter, "Institutional Discourse and Interlanguage PragmaticsResearch," we survey research in interlanguage pragmatics that is situated ininstitutional talk. We argue the main thesis of the book, namely that institutionaltalk is a rich setting for the investigation of interlanguage pragmatics, and thatsuch work/investigations can easily be accomplished in the framework ofinterlanguage pragmatics. We review 13 studies that were available at the timeof this writing that illustrate the benefits of the approach. We demonstrate thatsuch studies can and do address the typical concerns of interlanguage pragmaticanalysis such as the realization of a range of speech acts, while at the same timeaddressing concerns such as miscomrmmication, input, communicative success,and conversational strategies.

In her chapter, "Writing Center Interaction: Institutional Discourse and theRole of Peer Tutors," Williams studies 10 writing tutorials that take placebetween four writing tutors and five students at a university. This design allowseach student to be observed in two tutorials with different tutors. Stimulatedrecall interviews were also conducted. Williams reports that the input to the NEand L2 writers is not always the same. There are both quantitative and qualita-tive differences. Tutors produce more supportive interruptions of L2 writers, butmore advice for NS writers. Directives to nonnative speakers get more upgradersthan directives to native speakers. One possible explanation may be the effort onthe part of the tutors to facilitate L2 writer comprehension. This may also be afactor in the simpler and more explicit directive forms that the tutors use withthese students. Though characteristics of both dominance and solidarity arepresent in the writing center sessions, the data suggest that both participantswillingly take on nonreciprocal roles in their interaction. The balance is towardtutor dominance and authority, as shown through turn length, floor management,and their use of potentially face-threatening speech acts.

In the next chapter, Yates investigates the directives used by 18 teachertrainees doing their practice teaching in secondary schools in Australia. In"Negotiating an Institutional Identity: Individual Differences in NS and NNSTeacher Directives," Yates compares the directives produced by nine Australianbackground teachers and nine Chinese background teachers each in two fullclass sessions in a variety of school subjects that include ESL, business, math,music, science, and swimming. Teacher directives are compared by group(Australian and Chinese, male and female) and by individuals. Although, ingeneral, Australians employ a greater variety of mitigators than Chinesebackground teachers, the production data also shows that the Australian femaleteacher trainees use more mitigators than the Australian males and that theChinese background female teachers use more than the Chinese backgroundmales. In addition to gender and background, individual differences also play arole. Individuals vary in the way in which they project themselves throughlanguage use, in this case, in how they mitigate directives in the social identityof novice teacher. The variation found here is a useful reminder that all speakers


of a language, both native and normative, are actively engaged in projectingthemselves as individuals in relation to cultural norms with such resources asthey have at their disposal.

In chapter 4, Kerekes situates her study, "Before, During, and After theEvent: Getting the Job (or not) in an Employment Interview," in the office of anemployment agency that specializes in placing applicants in temporary employ-ment. Kerekes follows 48 job applicants, 24 NS and 24 NNS, also balanced forgender and type of job they seek, and 4 female middle-class staffing supervisorswho conducted the interviews that constitute the primary data of the investiga-tion. Kerekes also conducted background interviews with the staffing supervi-sors and debriefing interviews with both staffing supervisors and applicants aftereach employment interview. The outcomes of the interviews show that nonna-tive speakers have an equal chance of success, which in this case means gettinga referral for a job. Kerekes finds that comembership is likely to influence astaffing supervisor's choice of questions to the applicant as well as her moregenerous interpretations of the applicant's responses. This ultimately leads to thegreater likelihood of a placement in cases of high comembership. In this chapter,we see that, whereas clerical candidates have a higher success rate than theirlight industrial candidates, and female candidates are more successful than theirmale counterparts, there is no obvious advantage to being either a NS or NNS ofEnglish; neither was there an advantage for the NNSs in having a particularlyhigh level of L2 ability. As was recounted by the supervisors in their follow-upinterviews with Kerekes, successful encounters were related to sociolinguisticfactors such as the ability to demonstrate desirable employee characteristics,which proved equally possible for job candidates with a variety of languagebackgrounds.

Chapter 5, "Discourse Strategies in the Context of Crosscultural andInstitutional Talk: Uncovering Interlanguage Pragmatics in the UniversityClassroom" is set in a university physics lab. In the first part of their study,using an interaction between a Korean teaching assistant and an Americanundergraduate, Davies and Tyler present an interactive microanalysis of anegotiation about cheating. In the second part, they investigate the possiblesociolinguistic factors which might influence the nature of the negotiation. Inorder to do so, they utilize six Korean and three American judges, all of whomhave teaching experience, ranging from one semester to several years. They askthe judges to determine whether or not the exchange would be acceptable orexpected in their respective university settings. Both groups seem to haveschemas for how a teacher might deal with a student suspected of cheating, butneither group suggested the schema in evidence in the Korean teaching assis-tant's talk. This leads Davies and Tyler to conclude that L1 influence is not theprimary factor in shaping the negotiation. Rather, each crosscultural situation ispotentially a new context in the interface between cultural and linguisticsystems. They interpret the interaction a complex construction built of L1pragmatics and of a partial understanding of the target language pragmatics,

Introduction 5

filtered through his perception of the target culture, and further shaped byresources and constraints of particular institutional contexts.

In her chapter, "English for Specific Purposes and Intel-language Pragma-tics," Tarone outlines work based on genre analysis and introduces readersfamiliar with the interlanguage pragmatics research to the English for specificpurposes (ESP) literature, which is typically overlooked in interlanguagepragmatics. Research on ESP examines ways in which members of particulardiscourse communities use language varieties (genres) to communicate with oneanother in their pursuit of common goals. Pragmatic effects are often at thecenter of ESP research, which stresses the way discourse communities agree touse language. A discourse community is a group of individuals who share a setof common public goals, use mechanisms of intercommunication to provideinformation, use one or more genres to pursue their goals, use some speciallexis, and include members with content and discourse expertise. Fundamentalto this approach is the expert-novice distinction among members of a discoursecommunity. Particularly important for the study of the development of pragma-tics is that all newcomers or novice participants must learn the rules of thediscourse community, regardless of their status as native speakers or normativespeakers. In fact, a NNS can as easily be an expert in a discourse community asa NS. This would apply, for example, to the novice teachers in Yates' study,learning to become representatives of the institutions, and to the clients inKerekes' study learning to be successful job applicants. Both groups includedNS and NNS participants, all of whom sought to be successful within therelevant discourse community. We can see that the genre analysis challenges usto look beyond the assumption that the native-nonnative divide is the mostimportant one in learning pragmatics in a new setting.

Tara Gibbs illustrates genre analysis in the setting of a four star conferencehotel in her chapter, "Using Moves in the Opening Sequence to Identify Callersin Institutional Settings." Gibbs investigates the call-in, part of a class ofcommunicative events with the single goal of requesting another department todo something at the hotel. She demonstrates that novice housekeepers who usethe structure of a conversational opening are misidentified by other hotelemployees, whereas the experts who use the opening moves distinct to the call-in are immediately recognized as housekeepers. That is, the former are notviewed as members of the discourse community, whereas the latter are. Gibbsanalyzes the call-ins of one primary novice and one expert with examples drawnfrom other novices and experts. She compares the call-ins of novices beforeinstruction modeled on the expert call-in to the same novices' call-ins afterinstruction. Novices show successful outcomes after instruction that involvesmodeling; that is, they are appropriately identified by the other employees whomthey are required to call as part of their job. The use of expert models helpsnovices become more like the normative members of the discourse community.

In chapter 8, "Practical Considerations," Bardovi-Harlig and Hartford dis-cuss the fundamentals of getting started on a project involving institutional talk.As an adjunct to the material contained in the chapters of this volume, we asked


the authors of those chapters and other researchers who have written on institu-tional talk to share their practical advice and experiences with us. Their insightshelped us develop a set of guidelines to conducting research on institutional talk,which we illustrate with their first-hand accounts. These guidelines are intendedto help the researcher when setting up a project on institutional research. Someof the guidelines deal with the daily minutiae required to carry out such researchsuccessfully; details such as being thorough with checking equipment andlocateons, and listening/viewing recordings as soon as they are available. Otherguidelines are intended to help the researcher gain access to the institution andcarry out the work in such a way as to have things run as smoothly as possiblefor all concerned. They suggest ways to approach the institution, ways toaccommodate the potential participants as well as the institution itself; and theyaddress what still remains to be done after data gathering is complete. Amongother things, these latter guidelines contain some simple politeness suggestions,perhaps commonsensical for many. Sometimes, however, researchers forget topractice these niceties as they carry out the project, even though theyunconsciously do so in their daily lives. These guidelines serve as reminders.

We hope the reader will find that this volume serves as a handbook forinterlanguage pragmatic research utilizing one type of authentic data. Not onlydoes it contain practical suggestions for how to gather the data for such research,the chapters illustrate various ways in which the data might be approached inorder to address the research queries of the investigator. Close linguistic analysisof the discourse, consideration of sociolinguistically relevant factors such asgender, status, and L1 background, the relation of various outcomes to thediscourse observed, and practical applications of the project are all presented asavenues of investigation expanding the current practice in interlanguageresearch.

REFERENCE

Kasper, G. (1997). The role of pragmatics in language teacher education. In K. Bardovi-Harlig & B.S. Hartford (Eds.), Beyond methods: Components of language teacher education (pp. 113-136). New York: McGraw Hill.

1

Institutional Discourse and InterlanguagePragmatics Research

Kathleen Bardovi-HarligBeverly S. Hartford

Indiana University

Interlanguage pragmatics research investigates the acquisition of pragmaticknowledge in second languages, deriving its research methods from comparativecross-cultural studies and second language acquisition research. Both disciplinesplace a high value on the control of variables that facilitate comparison acrossspeakers, whether across cultures and languages, between native and nonnativespeakers, or among learners at different stages of acquisition. The orientation ofthese disciplines exerts a strong pull toward experimental data collectionprocedures. However, the fundamental nature of the very object of study-language use-argues for the study of situated authentic discourse. This tensionbetween the controlled and the authentic has been pointed out in a number ofdiscussions of research methods in interlanguage pragmatics research. AsKasper and Dahl (1991) observe:

IL pragmaticists are caught between a rock and a hard place. With the exceptionof highly routinized and standardized speech events, sufficient instances ofcross-linguistically and cross-culturally comparable data are difficult to collectthrough observation of authentic conversation. Conversely, tightly controlleddata elicitation techniques might well preclude access to precisely the kinds ofconversational and interpersonal phenomena that might shed light on thepragmatics of IL use and development. Clearly there is a need for moreauthentic data, collected in full context of the speech event, (p. 245)

The identification of the problem and the call for greater use of authenticconversation as data has done little to shift the balance toward the study ofauthentic talk in interlanguage pragmatics research, however. As we show in thefollowing discussion, tightly controlled elicitation tasks are still the norm.

The problem that we address in this chapter is how interlanguage pragma-tics research can retain the highly valued replication and comparability ofexperiments that have dominated interlanguage pragmatics studies and at thesame time meet the expressed desire of incorporating authentic data into theinterlanguage pragmatics corpus. The goal of this chapter is to introduce inter-language pragmatics researchers (and students of interlanguage pragmatics) toinstitutional talk and to demonstrate that it is a source of authentic discourse thathelps fulfill the expressed needs of the field. Institutional talk may be

7


understood as talk between an institutional representative and a client (e.g., afaculty advisor and a graduate student, or an interviewer at a job agency and anapplicant) or between members of the same institution (also called workplacetalk, such as talk between a nursing supervisor and a nurse, or among hotel orfactory employees). We return to a more technical definition in the next section.

We are well aware that some linguists in other areas need no introduction toinstitutional talk or other forms of conversation. In this chapter we addressourselves specifically to interlanguage pragmatics and interlanguage pragmaticsresearchers as we adopt and adapt this source of data as our own. We do notintend to provide a comprehensive overview of research on institutional talk, butrather to review recent interlanguage pragmatics studies that have investigatedinstitutional and workplace talk and to demonstrate their contribution to the fieldof interlanguage pragmatics. In the next section we provide a definition of andintroduction to institutional talk, and in the following section provide a briefsurvey of data collection in interlanguage pragmatics research.

INSTITUTIONAL DISCOURSE

Although there are varying definitions of institutional discourse in the field, wedraw upon those offered by Sarangi and Roberts (1999) and Drew and Heritage(1992). Sometimes referred to as workplace talk or institutional interaction,interaction, these authors agree, "is institutional insofar as participants'institutional or professional identities are somehow made relevant to the workactivities in which they are engaged" (Drew & Heritage, 1992, pp. 3-4). For ourpurposes this includes two main categories: interactions between institutionalrepresentatives and clients, and interactions between members of the institution.

A few of the former contexts, referred to as frontstage by Sarangi andRoberts (1999, p. 20) in which institutional discourse has been studied includesettings such as business, as in employment interviews (Akinnaso & Ajirotutu,1982; Gumperz, 1982a); legal, especially court testimonies (Conley & OBarr,1990; Gumperz, 1982b; Philips, 1990); educational, including school counselingsessions of students (Erickson & Schultz, 1982; He, 1994) and teacher-studentdiscourse (Mehan, 1994); and medical, including interviews between doctorsand patients (Fisher & Todd, 1983; Labov & Fanshel, 1977; Tannen & Wallat,1993).1

The second category, interaction that takes place among workers of theinstitutions rather than between client and institutional representative (backstagefor Sarangi & Roberts, 1999, p. 20), has been less widely studied. However,some work can be cited: ODonnell (1990) discusses the interaction in disputesbetween labor and management; Linde (1988) examines pilot and air trafficcontroller discourse; Erickson (1999) discusses physician—apprentice discourse;and Cook-Gumperz and Messerman (1999) study professional collaboration.

1 Drew and Heritage (1992) covers a wide range of these settings, as does Sarangi and Roberts(1999).

Institutional Discourse and ILP Research 9

Institutional discourse differs from ordinary conversations in three primaryways: goal orientation, constraints, and frameworks (Levinson, 1992). Drew andHeritage (1992, p. 22) summarize these facets of institutional discourse asfollows:

1. Institutional interaction involves an orientation by at least one of theparticipants to some core goal, task or identity (or set of them)conventionally associated with the institution in question. In short,institutional talk is normally informed by goal orientations of a relativelyrestricted conventional form.

2. Institutional interaction may often involve special and particularconstraints on what one or both of the participants will treat as allowablecontributions to the business at hand.

3. Institutional talk may be associated with inferential frameworks andprocedures that are particular to specific institutional contexts.

These three characteristics are those which make institutional discoursesuitable data for interlanguage pragmatics research: They contribute to thecomparability of multiple interactions. Whereas conversations do not tend tohave such constraints and are therefore not so easily comparable, institutionalinteractions often include expected norms of interaction such as turn-taking,constant social relations/roles, and asymmetrical power relationships.2 Sarangiand Roberts (1999) also list as additional constraints on institutional talk thefollowing: "decision-making and problem-solving; the production and regula-tion of professional knowledge and activities concerned with professionalcredibility, role relationships around issues of identity and authority" (p. 11).

TYPES OF DATA IN INTERLANGUAGEPRAGMATICS RESEARCH

Interlanguage pragmatics research has been slow—and even reluctant—toembrace conversational or other authentic data, as an inventory of the datacollection methods used in published volumes of interlanguage pragmaticsresearch shows. The survey of methods in use in interlanguage pragmatics doneby Kasper and Dahl in 1991 shows that out of 34 production studies, reporting35 data collection procedures, only 2 (6%) used authentic data, 19 (54%) useddiscourse completion tasks (DCTs), and 14 (40%) used role plays. InInterlanguage Pragmatics (Kasper & Blum-Kulka,1993), seven chapters reportdata collection. One uses a role play and five use some type of DCT (4 DCTsand one written dialogue construction). One uses authentic dinner tableconversations (Blum-Kulka & Sheffer, 1993). In Speech Acts across Cultures:Challenges to Communication in a Second Language (Gass & Neu, 1996), 11chapters report 13 data collection techniques. No chapters report spontaneous

2 Although most of this work investigates face-to-face oral discourse, there is some work on writteninstitutional discourse, particularly on business writing (Moran & Moran, 1985; Yli-Jokipii 1991).


conversation: One study collected TV commercials (Schmidt, Shimura, Wang,& Jeong, 1996), and one used talk generated in the course of a puzzle-solvingtask performed by dyads (Geis & Harlow, 1996). The remaining 11 usedcontrolled elicitation tasks: Five used DCTs (3 written, 2 oral), 4 used role plays,and 2 used retrospective interviews to collect reports of speech acts. Amonograph entitled Interlanguage Refusals (Gass & Houck, 1999), reports onrole plays in which participants play themselves (EFL students visiting theUnited States and staying with host families) faced with a number of undesirableoffers, invitations, or suggestions.

Although many readers will be familiar with the elicitation tasks used ininterlanguage pragmatics, we will briefly review the major types here. Produc-tion questionnaires are the most commonly used elicitation task in interlanguagepragmatics. In fact, Rose (2000, p. 1ll) characterizes the dominance ofproduction questionnaires as reflecting the field's "overwhelming reliance onone method of data collection." Production questionnaires include any of avariety of questionnaires that elicits speech act production data (Johnston,Kasper, & Ross, 1994, 1998; Rose, 1997). Among production questionnaires,the most common type is the DCT. DCTs, often described as written role plays,present a description of a situation (called a scenario) and ask the participant torespond. There are at least two types of DCTs: open questionnaires in which noturns are provided, and dialogue completion tasks in which an initiating turn or arejoinder is provided (Kasper, 1991). Production questionnaires have developedto include oral as well as written DCTs (Murphy & Neu, 1996; Yuan, 1998),computer-interactive DCTs where the computer gives and takes up to three turns(Kuha, 1997), the cartoon oral production task (COPT) designed for younger L2learners (Rose, 2000), and dialogue writing (Bergman & Kasper, 1993).

Two types of role plays are distinguished in the literature: closed and openrole plays (Gass & Houck, 1999; Kasper & Dahl, 1991). Closed role plays aremore commonly known as oral DCTs; in these a respondent replies to a scenarioprovided by the researcher on tape. Such oral DCTs do not offer the opportunityto study interaction. Open role plays (or simply, role plays) describe a scenarioprovided by the researcher which two or more participants act out. The outcomeof the interaction is often unspecified so that participants have some flexibilityin their contributions.

In contrast to responses to production questionnaires and role playscenarios, authentic discourse takes place without the instigation of a researcher.We note that authentic discourse may be either oral or written, and eithermonologic, dyadic, or multipartied. Conversational data constitutes the mostfamiliar form of authentic discourse and the one most generally referred to indiscussions of data collection in interlanguage pragmatics research. In thischapter we focus on oral dyadic exchanges that are the intended target of mostinterlanguage pragmatics research, but include one example of interactional e-mail institutional discourse that has been investigated in interlanguagepragmatics (Hartford & Bardovi-Harlig, 1996a).


The major data collection techniques have received numerous assessmentsin the literature (e.g., Cohen, 1996; Cohen & Olshtain, 1994; Gass & Houck,1999; Houck & Gass, 1996; Kasper & Dahl, 1991; Rose, 1997; Wolfson, 1989).We draw on these discussions to compare the resulting language samples bythree main features: comparability, interactivity, and consequentiality. Thesethree features reflect the articulated values of the field (Bardovi-Harlig, 1996;Gass & Houck, 1999; Kasper & Dahl, 1991, among others). Comparabilityassures that language samples can be reasonably compared. Comparability isseen to be the result of control in the collection of production data (cf. Kasper &Dahl, 1991). Interactivity characterizes language samples in which speakershave the opportunity to take turns. Consequentiality refers to the fact that thereis a real world outcome, or consequence, to naturally occurring talk, which oftenextends beyond the verbal exchange to establish itself in order to accomplishgoals (Drew & Heritage, 1992; Sarangi & Roberts, 1999).

In Table 1.1 we use these three features to characterize the data yielded bythe major collection techniques (i.e., production questionnaires, role plays,conversations, and institutional talk). We use the conventional notation ofindicating that a feature is present or absent, (e.g., [+/- comparable]), althoughwe note that this is a shorthand because these features likely represent continuarather than binary categories. Production questionnaires, by virtue of the factthat they are highly controlled tasks, yield language samples of high compara-bility. However, as widely noted, they are neither interactive, nor consequential.Role plays retain the experimental control of production questionnaires,resulting in comparable language samples, and they have the additionaladvantage of being interactive. Kasper and Dahl (1991) observe that open roleplays "represent oral production, full operation of the turn-taking mechanism,impromptu planning decisions contingent on interlocutor input, and hencenegotiation of global and local goals, including negotiation of meaning" (p.228). On the other hand, role play talk is not consequential; Gass and Houck(1999, p. 29) report, "Role plays are just that, role plays, with few, if any, real orreal life consequences." Aston (1993, p. 229) also notes that "it is difficult forthis technique to reproduce the interpersonal context of naturally occurringtalk."

An additional type of interactive language sample that is collected byexperimental techniques is the simulated task as shown in the second column ofTable 1.1. In simulated tasks speakers are brought together by researchers in anexperimental setting and asked to complete a task such as puzzle solving (Geis& Harlow, 1996) or teaching a computer program (Woken & Swales, 1989) forthe purpose of observing their talk in methods other than role play. Althoughsimulated tasks are rarely used in interlanguage pragmatics—of the volumesreviewed, the sole representative is the puzzle-solving task used by Geis andHarlow (1996)—they are familiar in studies of NS-NNS conversation (e.g.,Gass & Varonis, 1985; Selinker & Douglas, 1985; Woken & Swales, 1989;Zuengler, 1989). Like role enactments used by Gass and Houck (1999), butunlike some role plays, simulated tasks have the advantage that participants

12 Bardovi-Hartig and Hartford

Table 1.1. Analysis of Language Samples by Source

Productionquestionnaires

+ comparable- interactive- consequential

Role playssimulated tasks

+ comparable+ interactive- consequential

Conversation

- comparable+ interactive+ consequential

Institutional talk

+ comparable+ interactive+ consequential

speak as themselves rather than assuming a hypothetical role. Nevertheless,simulated tasks are generally undertaken by speakers with no social connection,in contrast to spontaneous conversations in which the speakers have some sociallink to their interlocutors.3 Finally, like role plays, simulated tasks lackconsequences beyond the task itself.

Moving to the third column of Table 1.1, conversation is characterized asboth interactive and consequential. Whereas role plays have few consequences,every conversation has socioaffective consequences. In real-world conversationsspeakers negotiate face, they take turns, and they attend to aspects such as topicand key (Hymes, 1964) in the process of being cooperative speakers (Grice,1975). Deviations from the accepted norms can result in changed evaluations bythe participants of each other, or in the need to reinterpret contributions. Thesemay be long-lasting consequences or fleeting; no matter which, participants inconversations are invested and therefore affected by any interaction. Suchconsequences have been frequently reported in the literature: Socioaffectiveconsequences were documented by Beebe and Cummings (1985, 1996) in theircorpus on refusals in telephone calls. In 1985, TESOL was held in New Yorkand the local organizing committee called local members to volunteer. However,TESOL had been scheduled during Passover, a major Jewish holiday and manyTESOL members declined to volunteer during the holiday. As Beebe andCummings observed, "It is extremely unlikely that a hypothetical situation couldevoke such strong emotion as the actual scheduling of TESOL during Passoverin 1985" (1996, p. 79). A second example of socioaffective consequence isfound in research on conversational openings and closings. Research on conver-sational openings reports that when speakers engage in these openings, they (re-)establish their relationship with their interlocutors, reveal information about thestate of the relationship, establish status relations, and ascertain whether aninteraction will be friendly or not (Irvine, 1974). Closings on the other hand, notonly serve to organize the termination of a conversation, but they also reinforcerelationships and support future interactions as they anticipate lack of access toeach other after the closing (Button, 1987; Goffman, 1971; Hartford & Bardovi-

3 Note that the lack of social connection between speakers in simulated tasks provides a counterpointto Aston's (1993) observation that real-world relationships between actors in role plays who areassigned other roles "might constitute an additional, and unplanned, influence on the role play" (p.229). The generalization that we draw from these two observations is that relationships that areinvented or imposed on participants add another layer of complexity to the interpretation of talk thatis not present when authentic relationships are in play.


Harlig, 1992; Omar, 1992; Schegloff & Sacks, 1973). Such work on relation-ships cannot be observed in simulated tasks.

The advantages of studying conversation for interlanguage pragmaticsresearch derive from the fact that conversation is a natural occurrence, as notedabove: It is spontaneous authentic language use by speakers who are speaking asthemselves, in genuine situations, with socioaffective consequences. In addition,conversations exhibit a variety of speech acts and attributes in a single encounterand constitute a rich source of data. From the interlanguage pragmatics perspec-tive, however, conversational data is seen as being insufficient from the stand-point of comparability of the language sample. Except for the most routine andsuperficial exchanges, conversational samples are not viewed as beingcomparable and role plays are seen as more desirable for interlanguage pragma-tics research with respect to that feature: "Role plays have the advantage overauthentic conversation in that they are replicable" (Kasper & Dahl, 1991, p.229). In fact, many reviews of interlanguage pragmatics research design do notinclude conversation and only compare experimental methods. Furthermore, it isgenerally accepted that observation does not yield enough tokens of a particularattribute to constitute a reasonably sized language sample (Wolfson, 1986,1989).

Thus far in the literature, there has been a perceived trade-off betweencomparability and consequence. However, as the last column of Table 1.1shows, there is at least one type of talk which is authentic and consequential, andat the same time can be compared to many other samples taken from the samesetting: institutional talk.

The goal-oriented nature of institutional talk allows for comparison acrossspeakers in the same way that experimental interactions allow researchers tocompare speakers (e.g., Selinker & Douglas, 1988; Woken & Swales, 1989;Zuengler, 1989). Comparison is facilitated because similar conversations recurboth with the same participants (the institutional representatives) and withdifferent participants (the clients), in similar settings, and on similar topics.Furthermore, as Agar (1985) and Erickson and Schulz (1982) have noted, manyinstitutional encounters have very similar structures, which are quite differentfrom informal, spontaneous conversations. Within one type of interaction,similar time frames are maintained: For example, writing tutorials last about anhour; advising sessions are scheduled in 15-minute intervals. Clients andinstitutional representatives may draw on their shared knowledge of the structureand goals of institutional events to assure success, as Cameron and Williams(1997) show in interviews between a NNS nurse and her NS patients. Theinstitutional setting makes it possible to control for some of the variables that aretroublesome when working with spontaneous events, and makes it possible tocompare results.

An example regarding comparability of institutional encounters is found inthe closings of advising sessions studied by Hartford and Bardovi-Harlig (1992).Nonnative speakers sometimes had difficulty closing the conversations. Since allsessions had a similar structure and known time frame (about 15 minutes), this


shared information about the session might serve as background knowledge,which all participants could utilize and assume of each other. Additionally,because faculty advisors participated in multiple sessions, they served asconstants: It was possible to examine how they interacted in both successful andless successful closings with both native speakers and nonnative speakers.Native speaker students could be compared to the normative speaker students inorder to determine what the similarities and differences were across subjects andspeech events, and thus discover how the unsuccessful closings differed fromthe successful ones. Kasper and Dahl (1991) observed that the advising sessionsyielded authentic language samples that could be compared with each other "asNSs and NNSs interacted with the same interlocutors and in the same statusrelationship, comparability of the data was ensured" (p. 231).4

In the following section we demonstrate the benefits of using institutionaltalk for interlanguage pragmatics research.

INSTITUTIONAL TALK AS A SOURCE OF AUTHENTIC DISCOURSEFOR INTERLANGUAGE PRAGMATICS RESEARCH

In this section we discuss the advantages of utilizing institutional talk as asource of spontaneous discourse in interlanguage pragmatics research. In the last10 years there have been a small but significant number of studies that haveexamined interlanguage pragmatics through institutional talk and that illustrateits benefits as a tool for the study of interlanguage pragmatics. Data collectedfrom institutional talk offer a good balance of authenticity and consequence onthe one hand and potentially large data sets resulting from predictable occur-rences and control of variables on the other. Table 1.2 summarizes interlanguagepragmatics studies of institutional talk that takes place between an institutionalrepresentative and a client.5 Table 1.3 summarizes the smaller set of interlan-guage pragmatics studies that investigate talk between institutional members.One study appears on both tables: Cameron and Williams (1997) report on anurse in training who is a normative speaker of English interviewing twopatients in a hospital psychiatric unit (institutional talk) and talking to hernursing supervisor (workplace talk).

As Tables 1.2 and 1.3 show, authentic discourse may be studied frommultiple perspectives and a variety of analytic frameworks. To date, the theore-tical framework that has dominated interlanguage pragmatics has been speechact analysis, and thus it is also a common—but not the exclusive—method ofanalysis of institutional talk. Perhaps the most important point we can make hereis that interactive data are not restricted to any one type of analysis, including

4 They also note, however, that the advising sessions might be "confronting the limits of cross-cultural comparability" (p. 231) because institutions and hence the talk might be culture specific. Wereturn to this point later in the chapter.5 In Table 1.2 we include only studies that investigate discourse directly. Studies based on interviewsin which learners and their interlocutors report on speech behaviors in institutional settings, such asLi (2000), have not been included.


speech acts. Although this argument is beyond the scope of this chapter, it couldbe that once researchers see the richness of authentic discourse and possibilitiesfor investigation, the dominance of the speech act model might naturally lessen.6

Methodological Considerations:Institutional Talk as a Natural Experiment

The advantages of studying authentic discourse are often sacrificed because ofthe perceived difficulty of data collection. Unpredictability of occurrence, lackof control of variables (Beebe, 1994; Beebe & Cummings, 1996), and paucity ofsufficient features in a sample (Wolfson, 1986) are often cited as methodologicalproblems of studying authentic data. In addition, advantages of experimentaldata (to the exclusion of authentic data) have included the availability ofretrospective interviews (Cohen, 1996; Cohen & Olshtain, 1994). In this sectionwe begin by investigating these methodological points in order to argue thatinstitutional talk yields authentic language samples that can inform interlan-guage pragmatics research. In the following sections we move on to theadditional characteristics of language use and development that can be studied insuch data.

Institutional talk usually takes place in relatively fixed locations within theinstitution, and this aids in data collection. For example, advising sessions occurin the advisors' offices, writing tutorials occur at the writing center, and psychia-tric interviews take place at the hospital. Because institutional talk takes place inpredictable locations and often at predictable times through the scheduling ofappointments or the hours of operation, offices and other settings can beprepared for data collection before the speech event. In the larger settings of theworkplace, Clyne (1994) identified areas within factories that were sites ofpredictable talk and were relatively quiet enough for recording. We should alsonote that there are some types of institutional encounters which change location,but which could still be investigated. Sarangi and Roberts (1999, p. 5) point outthat workplaces may vary, and in some cases the institutional representative maygo to the client, as do, for example, nurses who deliver homecare. In this case,the actual physical setting will vary, but still, the essential psychological settingdoes not, and so the nature of this discourse is much the same, that is,institutional.

Moreover, in addition to the sites of occurrence of the institutional talkitself, the nature of institutional talk is such that discourse topics and attributesof talk can be anticipated in advance of the data collection. Because of the goal-oriented nature of institutional discourse, specific speech acts can be anticipatedin the course of the speech event. Thonus' (1999) data on writing tutorials

6 Readers who are interested in analyses other than or in addition to the speech act framework shouldnote Cameron and Williams, 1997; Gass and Houck, 1999; Hartford and Bardovi-Harlig, 1992;Kerekes, 2003; Tarone and Kuehn, 2000, Tyler, 1995, and Tyler and Davies, 1990, and manychapters in this volume.

Table 1.2. ILP Studies of Institutional Discourse

Study

Bardovi-Harlig &Hartford (1990, 1993a)

Tyler &Davies( 1990)

Hartford & Bardovi-Harlig (1992)

Ellis (1992)

Tyler (1995)

Bardovi-Harlig &Hartford (1996)

Hartford & Bardovi-Harlig (1996a)

Cameron & Williams(1997)

Thonus(1998, 1999)

Churchill (1999)

Kerekes (2001)

Yates (2000)

Focus

Suggestions and Rejections(refusals)

Communication missteps

Closings

Requests

Miscommunication

Input

Requests

Communicative success

Directives

Requests

Linguistic behaviors andconversational strategies

Directives (mitigation)

Institutional representative

NS

NNS teaching assistant

NS

NS teachers

1 NNS internationalteaching assistant

NS

2 NS faculty members

1 NNS nurse

NS writing tutors

NS and NNS teachers

4 NS staffing supervisors

9 NNS and 9 NS studentteachers

Client

NS and NNS graduatestudents

NS undergraduate student


2 child learners of ESL

NS student


NS and NNS students

2 NS patients

NS and NNSundergraduates

Learners of EFL in partial-immersion

48 applicants 22 NS,22NNS, 4 bilinguals

Students

Institutional setting

Academic advisingsessions

Physics lab


Classroom discourse

Tutoring session


E-mail messages to faculty

Hospital psychiatric unit

Writing tutorials

Academic setting

Employment interview

Classroom teaching

Table 1.3. ILP Studies of Workplace Discourse

Study

Clyne (1994)Main corpus, 150 hours.Secondary corpus,32 hours (meetings)

Cameron & Williams(1997)

Focus

Speech acts (complaints/whinges, directives,commissives, apologies),turn-taking, small talk

Communicative success

Institutional employee

37 NNS employees insupervisorial, equal, andsubordinate positions.Secondary corpus combinedw/primary, 39 NNS

1 NNS nurse

Workplace interlocutor

NNS employees insupervisorial, equal, andsubordinate positions.Secondary corpus, NS andNNS interlocutors at meetings

Nurse's clinical supervisor

Institutional setting

Main corpus,Nipponcar, Amcar,Weaver, Elektro,Catering.Secondary corpus,Weavers, EmploymentOffice, Parents Group

Hospital psychiatric unit


included problem statements by the students and suggestions by the tutors.Students make requests that help them complete their assignments in the courseof being students, whether they be elementary school-aged children (Ellis, 1992)or teenagers in high school (Churchill, 1999). Another such example is theoccurrence of directives in the talk of student-teachers addressed to secondaryschool pupils (Yates, 2000).

Participants in an institution can generally be identified, and this allowsresearchers to describe speakers according to the variables relevant to a givenstudy. The institutional speaker is generally known to the researcher, andinformation about the institutional client can be ascertained. Moreover, the rolesof the participants are institutionally proscribed and the relationships aregenerally fixed. In her study of writing tutorials, Thonus (1998) balanced theinstitutional representatives by selecting 12 tutors from among the 20 employedat the Writing Tutorial Services, including six tutors who met with studentswhose paper topic was the same as their area of expertise and six tutors who metwith students whose paper topic was not. She further balanced the client-participants so that seven of the tutorials took place between students and tutorswho had met with each other before, and five tutorials took place withparticipants who had not yet met their tutors. In this way, Thonus was able tocompare the tutorials for familiarity with topic on the one hand and familiaritybetween the participants on the other. Only tutorials for which studentspresented drafts of their papers were selected for inclusion in the study, whichsomewhat restricted the stage of the writing process at which the student soughtassistance, thus further assuring the similarity of the tutorial sessions. Thisillustrates one way institutional roles and attributes can be used to balanceparticipants in a study.

The concentration of occurrences of talk particular to given institutionsleads naturally to a corpus with a substantial number of tokens for analysis. Forexample, Yates' (2000) study of student-teachers yielded 2,973 directives (1,527from 9 Anglo-Australian teachers in 13 hours, 18 minutes of classroomrecordings, and 1,446 from 9 Chinese background NNS teachers in 12 hours and18 minutes of recording). Thonus' (1998) study of 12 writing tutorials yielded441 directives addressed to NS students and 343 to NNS students. The size ofthe corpora—as realized by number of tokens of the targeted speech act—allowed both Thonus and Yates to establish pragmatic profiles for each of theindividual participants. Understanding differences among speakers rather thanreporting group data is a necessary step to understanding acceptable usage andvariation in the target language by both native speakers and normative speakers.

Two longitudinal studies of the use of requests in a school setting alsoresulted in good-sized corpora. Ellis (1992) recorded 410 requests made by twochildren, and Churchill (1999) reported 184 requests of 37 participants. It isworth noting that even within the institutional setting, certain individuals maynot produce the target speech acts, even if the majority of the participants do,and this, too, is revealing. Churchill reported that only 27 of the 37 participantsin his study produced requests. The nonproduction of requests by 10 learners


reveals a type of contextualized opting out not available in experimental tasksand perhaps not obvious in conversations where comparability amongconversations is hard to establish.

Cohen (1996) and Cohen and Olshtain (1993, 1994) advocate the use ofretrospective interviews to provide insight into production and perception in thearea of L2 pragmatics. Retrospective interviews constitute one type of externalevidence on which analysts may draw (Tyler & Davies, 1990) and have beenused successfully in the study of institutional talk (Erickson & Shultz, 1982, fornative speakers; for work in interlanguage pragmatics, Clyne, 1994; Kerekes,2001, 2003; Thonus, 1998; Tyler, 1995; Tyler & Davies, 1990) as well as inexperimental studies (e.g., Cohen & Olshtain, 1993).

In the tutoring session investigated by Tyler (1995), retrospective inter-views with both participants reveal insights into the motivation behind severalturns in which the student and tutor each intended to establish expertise regard-ing the basic knowledge that underpins the assignment under discussion. Thestudent had to write a computer program that could score a bowling game. Whenthe student asked "Do you know how to score the game?" and the tutoranswered "Yeah approximately", we learn from the student that she interpretedthat to mean that the tutor was not fully knowledgeable about scoringprocedures. We learn from the tutor that he believes he has given an unequivocalstatement of expertise. In order to remain modest he added a tag "approximate-ly." He asserted that any Korean would understand that his response meant,"Yes, I know how to score bowling." The interpretation of this turn proved to beabsolutely crucial to the tutoring session, and the session never fully recoversfrom the misinterpretation by both parties.

Related to retrospective interviews of the participants is the use of nonpar-ticipant consultants to interpret the data of others. An additional benefit of theinstitutional setting is the potential availability of consultants to interpret thedata of others. This may prove to be extremely valuable in areas in which theanalysts do not have subject expertise or institutional membership. (See Selinker& Douglas, 1989 and Douglas & Selinker, 1994, on the use of the subject-specialist informant in their work on discourse domains.)

The judgments or interpretations of consultants provide a second type ofexternal evidence for an analyst. Tyler and Davies (1990) studied a conversationbetween a Korean teaching assistant and an American undergraduate involving anegotiation over poor grades. Tyler and Davies enlisted the help of Americanundergraduates and two Korean graduate students to interpret the conversationbetween the teaching assistant and the undergraduate. Both sets of consultantsprovided insightful comments on both the structure and the content of theinteraction. The American undergraduate consultants' responses confirmed theresearchers' own expectations for the structure of a negative evaluativeexchange (i.e., why the student had earned a low grade): They expected theteaching assistant's first turn to provide a general overview of the problem orrefer to the most important error. In contrast, the Korean graduate-studentconsultants felt that the strategy used by the Korean teaching assistant,


beginning the explanation of the poor grade by referring to the more minorerrors and building to the main one (called the "inductive/collaborativestrategy," p. 401), is less threatening and more face-saving than the strategyexpected by the Americans. However, the Korean teaching assistant's strategyled many American undergraduate consultants to conclude that the teachingassistant did not really know why he had given the poor grade he had. Incontrast, in his retrospective interview, the teaching assistant reported that heknew that the student was upset and had wanted to soften the effect of hisstatement concerning the student's failure. The American consultants rejectedthe sympathetic interpretation.

A second study to use consultants was a study of requests which were madeof faculty by students via e-mail (Hartford & Bardovi-Harlig, 1996a).7 Therequests were presented to faculty consultants who rated the degree ofimposition of the content of each request. There was a high degree of agreementon the level of imposition among the faculty. The use of such consultants helpedconfirm the validity of the analysis. In a follow-up to the study the samequestionnaire about level of imposition was administered to domestic andinternational (NS-NNS) graduate students in the same department as the faculty.The results of the graduate student survey revealed that the students did notseem to have the same perception as faculty concerning the imposition of therequests, particularly for some of the high imposition categories. Requests thatstudents assumed were simply a part of the faculty member's job were generallyseen as low imposition, even if they required extra time and effort on the part ofthe faculty member. Using such students in a consultancy role helped to explainsome of the lack of politeness markers in e-mail requests: Often they werelacking just where the faculty and the students perceived the impositiondifferently.

As we have shown, the investigation of institutional talk addresses many ofthe methodological issues that researchers have raised and discussed in theinterlanguage pragmatics literature. However, the fact that studying institutionaltalk offers the advantage of studying comparable and authentic discourse is onlypart of what the study of institutional talk offers interlanguage pragmaticsresearch. The study of institutional talk opens the possibilities of studying morefeatures of language, language use, and language acquisition than have generallybeen considered in interlanguage pragmatics previously.

What We Can Learn From Institutional Talk

In this section we discuss how the study of institutional talk leads to an enrichedunderstanding of interlanguage pragmatics. We examine five main areas:understanding language in context, interaction of multiple features, timing andescalation, understanding goals and outcomes of talk, and acquisition ofpragmatics and the institutional rules themselves.

7 We realize e-mail is not oral like the other studies discussed. However, it is interactional andnatural. We include it here in order to discuss the use of consultants.


Understanding Language in Context. Understanding the role of context iscrucial to understanding language use and development. In controlled elicitationtasks such as production questionnaires and role plays, context translates intodeveloping credible and culturally appropriate scenarios to which speaker-writers respond (Rose & Ono, 1995). However, many scenarios seem to presenta rather simplified context and the attempts to deliver subtle differences in thedetails in a verbal description often meet with respondent fatigue, inattention, orinsufficient language proficiency (Bardovi-Harlig, 1999a). Attempts to over-come the limits of verbal delivery include the use of drawings (Rose, 2000),photographs (Rodriquez, 2001), and film (Bardovi-Harlig & Dornyei, 1998;Leary, 1994; Rose, 1997). Native speakers seem to need less fully specifiedscenarios than do nonnative speakers who are often not as familiar with certainsettings and rely more heavily on details of the scenarios (Bardovi-Harlig &Hartford, 1993b). Regardless of the scenario delivery however, individualscenarios (i.e., contexts) continue to emerge in study after study as an importantvariable (Bergman & Kasper, 1993). No matter how carefully a scenario ispresented to respondents in an experimental task, context will remain anunderexplored variable because analysts can never be sure that respondentsrecognize the setting presented to them. Instead, the study of authentic discoursecaptures speakers as they engage in meaningful interactions. The results ofrecent studies suggest that speakers, and especially learners, who have notexperienced interactions in settings presented in experimental tasks (such asattending school in the host environment, living in a dorm, or going to cafes),are more likely to guess at a response rather than to produce one based on otherexperience which might be relevant (Garton, 2000; Rodriguez, 2001). Inauthentic interaction, conversational as well as institutional, speakers areobserved in settings in which they have opted to interact. Although someinstitutional interactants may be more familiar with the institution than others,the fact of the matter is that, in an authentic institutional interaction, it isnecessary that they participate and achieve some outcome. Thus, even if they arenot fully versed in the norms of that particular institution's interactions,outcomes are arrived at through negotiation and feedback, which may in turncontribute to the acquisition of those norms by the participants.

The range of institutional settings provides a broad context for theinvestigation of institutional talk. Moreover, any single setting can provide arich background for investigation. Medical interviews, for example, occurwhenever patients come to medical establishments seeking health-related care.Such interviews may be between doctors and patients or between nurses orreceptionists and patients. In addition, talk between institutional members iscommon in this setting: talk between doctors and nurses, between nurses andsupervisors, and between nurses, to name only a few interactions. Cameron andWilliams (1997) studied the interactions between an NNS nurse and hersupervisor (talk between institutional members), which occurs naturally in asupervisory setting, and interactions between the nurse and two patients (talkbetween an institutional representative and her clients). Other general contexts


for institutional talk that have been studied in interlanguage pragmatics includewriting tutorials, employment interviews, and social service interviews (Tables1.2 and 1.3). Like the scenarios developed for controlled elicitation tasks, thecontexts within an institution may shift subtly to allow for the investigation ofvariables. In addition to the most obvious variables in interlanguage pragmaticsstudies—the differences between native and normative speaker, and differencesin L2 proficiency (not to mention the commonly investigated factors of age andstatus)—the influence of context-based variables can be compared. For example,Kerekes (2001, 2003) examines the difference in talk and resolution of mis-communication when interviewers regard job candidates as being trustworthyand when they do not. Thonus (1998, 2002) explores the characteristics ofwriting tutorial sessions that tutors and students rated as successful and thosethat they rated as unsuccessful.

A second type of context that is offered by the study of institutional talk isfound in the talk itself. The contextualization of turns at talk is not available indata from production questionnaires. Sequential contributions and negotiationscan be observed in role plays but these are not located in a larger authenticcontext. Taking a speech act perspective as an example, the definable contextsallow us to understand how speech acts are realized, how they are elaboratedand negotiated, and how and why they eventually reach a resolution between thespeakers. One promising area of investigation is the difference in speechaddressed to native- and normative-speaker clients by institutional represent-tatives (Tarone & Kuehn, 2000). Tarone and Kuehn observe that financialworkers may word questions differently with native speakers than with languagelearners, thus changing the interaction locally at the level of the turn for thelanguage learner. Although they point out that they could not follow that line ofinvestigation because different financial workers interviewed the native speakersand the language learner, this is an important point of comparison that couldeven be addressed by analyzing the pivotal turns of institutional representativesin existing larger corpora (e.g., Bardovi-Harlig & Hartford, 1993a; Thonus,1998, 1999).

We make no claim that these contexts are simple; in fact, they can be quitecomplex. On the other hand, institutional contexts are identifiable, and in spiteof their complexity, they are describable. More importantly for interlanguagepragmatics research, these contexts are replicable within an institution resultingin a background against which it is possible to study other variables such as age,gender, education, native language, language proficiency, and severity of theproblem presented to the institution by the client or to the client by theinstitution.

Multiple Pragmatic Features and Their Interrelations. Like conversa-tions, institutional talk also exhibits a number of structures in any one encounter.In medical settings, for example, institutional representatives ask questions, givedirections, and make diagnoses. Clients may also ask questions, and in additionthey are responsible for answering questions and acknowledging directions. The


corpus of talk from academic advising sessions alone has been a rich source ofdata for the investigation of many aspects of discourse including the construc-tion of student identity (Bardovi-Harlig & Hartford, 1993c), conversationalstructure and closings (Hartford & Bardovi-Harlig, 1992), and numerous speechacts that have included suggestions (Bardovi-Harlig & Hartford, 1990, 1993a),refusals (Bardovi-Harlig & Hartford, 1991, 1993a), and requests for advice(Bardovi-Harlig & Hartford, 1990). Similarly, in a single conversation betweeninstitutional members, Clyne (1994) identifies complaints, apologies, directives,and commissives.

In addition, a complete, multifaceted interaction allows researchers tounderstand particular exchanges in light of subsequent ones. For example, onlineinterpretations can be readjusted if information that occurs later in a conversa-tion demands it. In the advising interviews, Bardovi-Harlig and Hartford (1993a)showed that there were times when the advisor had understood studentresponses to advice on course schedules to be agreements at the time that theywere negotiated; however, later in the interviews the students indicated that theyhad not actually agreed to those suggestions, but were only giving what theybelieved were polite backchannels, and so schedule renegotiations had to takeplace. Thus, occurrences of multiple features within any single institutionalinteraction not only provide a potentially rich source of data, but may alsoprovide a view of the participants' interpretations of the interaction.

Timing and Escalation. There are often multiple occurrences of the samefeatures within a single encounter of institutional talk. This facilitates the studyof both the features and the effect of timing. For example, the talk of student-teachers not only contained directives to students (predictably), it also containedmultiple directives addressing the same issue (Yates, 2000). Thus, the institu-tional sample yields important information on escalation resulting from repeatedrequests. Yates found that in the classroom, the simple imperative is theunmarked form of a routine directive. In a series of directives given to a classwhich was waiting outside a room at the beginning of a lesson, a teacher said,"So come in quietly. Quietly/ Year seven." As they got more unruly, with loudtalking and jostling, the teacher switched to a more polite, more deferentialform, increasing the distance between herself and her students saying, "Excuseme Year seven/ could you go back outside and line up please" (p. 134). Yates'sdata also show how a series of requests can escalate over a class period (incontrast to directives issued in direct succession as in the previous example),moving from two uses of an unmarked imperative (" R/ go on with your work"and "S/ turn around"), to a warning ("If you don't stay X and do your work/ I'mgoing to have to send you out the back to talk to Mr. X, /alright?/ and do yourwork out there"), to the discipline phase delivered with a question abilitystrategy and a requestive politeness marker please, "R and S/ can both of youplease take your book/ and go and speak to Mr. X in the back room" (p. 134).


Goals and Outcomes. Hymes (1964) distinguishes between goals andoutcomes in interactive language speech events. These are particularly relevantfor institutional language. In such speech events, there may be conventionalgoals and conventionally expected outcomes. For example, in a consultation in adoctor's office, the conventional goal and expected outcome for bothparticipants is diagnosis of the patient. A further expectation, or goal, on the partof the patient participant is generally advice from the physician regardingtreatment. However, such expectations are not always met. It may be the casethat the physician cannot offer treatment advice, or even that the advice iscontrary to what the patient expects. In this case, the goals of the patient may notbe met, although there is certainly an outcome to the speech event. Thus goalsand outcomes may not always be the same, or may differ for each participant.There is then, in institutional talk, usually some real world, expected, tangiblegain to at least one of the participants, seen as the purpose of the conversationitself. In the institutional encounters studied in the interlanguage pragmaticsliterature, goals have included receiving a signed registration form in anadvising session, help with a writing assignment in a writing tutorial, aplacement in a job interview, financial assistance in a social service interview, ora psychiatric assessment in a psychiatric evaluation.

Comparing goals to outcomes affords one independent measure of successof the interaction. Goals are often stated by clients during the course of theinstitutional speech event and outcomes are generally observable. Tyler (1995)reports on participant goals in a tutoring session in which the student hadvoluntarily sought out the tutor for help on a major computer programmingassignment in order to increase her chances of success. Tyler reports that thestudent "genuinely needs to successfully complete the assignment" (p. 139). Thetutor, an international graduate student in Computer and Information Science,was enrolled in an advanced, elective English oral skills course. Graduatestudents enrolled in the course offered free tutoring to U.S. undergraduates aspart of the course. The graduate student tutor knew that all tutoring sessionswere videotaped and reviewed with the students by the professor. The tutor alsodesired to do his best, be evaluated favorably by the instructor, and improve hisEnglish skills. The outcome was quite different, however. Both participantsfound the session so unsatisfactory that each complained to the supervisor aboutthe other.

Another feature of institutional talk is that the outcomes are often recordedin written form. This is so common that Agar (1985) identified report writing asthe third and final stage of the institutional interview. There were manyexamples of written documentation in the studies reviewed here. The advisors inthe academic interviews (Bardovi-Harlig & Hartford, 1990, 1993a) entered theagreed-upon classes on a registration ticket and in the student's departmentalfile. The interviewers for the temporary job agency filled in an action documentthat recorded the interviewers decision on whether the job candidate should behired (Kerekes, 2003). In the writing tutorial, the tutor filled in both anassignment sheet and a session evaluation (Thonus, 1998). In the psychiatric


interview, the nurse took notes to which she later referred in meeting with hersupervisor (Cameron & Williams, 1997). In the case of the tutoring sessiondescribed earlier, an oral report in the form of a complaint was registered byboth of the participants (Tyler, 1995). The existence of such documentationprovides an additional external record of the outcome of the encounter—at leastas far as the institutional representative understood it.8

Observing institutional outcomes may be one way to begin to understandwhat pragmatic features are particularly significant in particular types ofinteractions. Kasper (1997, 1998) discusses the importance of determiningwhether divergence from the pragmatic norm is problematic or not, whereproblematic may include miscommunication as well as resultant negative socialperceptions. As Kasper (1998) notes, "pragmatic divergence itself is notproblematic if the social values indexed are acceptable or perhaps even valuedby the recipient" (p. 198). Regarding the outcome of the tutoring sessionbetween the Korean tutor and the undergraduate (Tyler, 1995), the interactionand negotiation of face were so unsatisfactory that both participants complainedabout the other to the tutor supervisor. In contrast, Cameron and Williams(1997) report that the outcome of the consultations between the nurse and hersupervisor were more satisfactory because of their mutual cooperation: Thenurse and her supervisor drew on professional knowledge and goals to assuresuccessful communication in spite of the limited linguistic competence of theNNS nurse. In studies of multiple speakers, mixed outcomes were reported.Student and tutor participants evaluated the success of the sessions (Thonus,1998), and in employment interviews NNS showed equal or better rates of beinghired than native speaker applicants (Kerekes, 2001).

The Study of Acquisition. Interlanguage pragmatics has been defined asthe intersection of the study of pragmatics and the study of second languageacquisition (Kasper, 1998). The study of institutional talk allows researchers theopportunity to study contextualized linguistic and pragmatic development on theone hand and the acquisition of institutional rules on the other. Neither area hasbeen fully explored within interlanguage pragmatics or studies of institutionaltalk, but both are promising areas of investigation.

One area of investigation is the acquisition of second-language pragmaticand linguistic features. As research interlanguage pragmatics increases itsemphasis on acquisition, it will be necessary to involve learners at differentlevels of proficiency (Bardovi-Harlig, 1999a, 1999b; Kasper & Schmidt, 1996:Rose, 2000). Thus far, interlanguage pragmatics has concentrated on advancedlevel learners due at least in part to the fact that many of the controlled tasksfavor higher proficiency learners. Whereas special tasks can be developed forlower level learners (see Rose, 2000 for an example), such a practice is not yet

8 Two of the studies (Kerekes, 2003; Tarone & Kuehn, 2000) examined encounters with another typeof written record, the application. The applications are filled out by the job or social serviceapplicants prior to the encounter and feature prominently throughout the interview.


in place. Because second language learners of all levels of proficiency interactwith institutions both as clients and as institutional members, the study ofinstitutional talk expands the study of language use to include lower proficiencyspeakers both as clients or members of institutions. Academic settings areparticularly good sources for learners at various levels of grammatical andpragmatic proficiency, as illustrated by Ellis's (1992) longitudinal study ofclassroom requests and by Bardovi-Harlig and Hartford's cross-sectional studyof learner talk during advising sessions in an academic intensive ESL program(Bardovi-Harlig & Hartford, 1997; Hartford & Bardovi-Harlig, 1996b).Churchill's (1999) longitudinal study of institutionally situated requests isnoteworthy because it does not take place in the host environment, but rather inan EFL setting, in a partial immersion high school in Japan in which English isused outside of class. Whereas academic settings often include language learnersranging from relatively low proficiency to higher proficiency, interactionsoutside academic settings are by no means limited to higher proficiency learners,as can be seen in the case of both clients (Tarone & Kuehn, 2000) andinstitutional representatives (Cameron & Williams, 1997).

An additional type of development that can be observed in the institutionalsetting is the learning of the institutional rules and the type of talk, or genre(Tarone & Kuehn, 2000), that is associated with the institution. Kasper and Dahlnote that the advising sessions might be "confronting the limits of cross-culturalcomparability" (1991, p. 231) because the practice of academic advising andhence the talk associated with it might be culture specific. To the extent thatinstitutions have no equal in other cultures, this creates a natural laboratory inwhich the acquisition studies of interlanguage pragmatics can observe how L2speakers learn new cultural-institutional rules in addition to learning thelanguage that goes with them. Moreover, nonnative speakers are not alone inhaving to learn institutional rules and institutional talk. Erickson and Schultz(1982) observed that academic counselors taught the interview to native-speakerstudents during the sessions themselves. (See also Tarone, chapter 6, thisvolume.) Bardovi-Harlig and Hartford (1993a, 1996), Tarone and Kuehn (2000),and Li (2000) observe the fact that new participants must learn the rules of theinstitution. Tarone and Kuehn distinguish between novice and expert clients.The social service interviews studied by Tarone and Kuehn, like the advisingsessions studied by Bardovi-Harlig and Hartford, are normally private and thus"the only way for a novice to become an expert is by participating in aninterview" (Tarone & Kuehn, 2000, p. 100). The fact that institutional talk has tobe learned by all clients—native speakers and normative speakers alike—affordsus the opportunity to observe the learning processes of native speakers as well asthose of normative speakers in institutions where both are clients and/ormembers. This includes settings such as graduate students enrolled in a particu-lar graduate program (Bardovi-Harlig & Hartford, 1990, 1993a, 1996; Hartford& Bardovi-Harlig, 1992, 1996a), social service applicants (Tarone & Kuehn,2000), and job applicants at temporary staffing agencies (Kerekes, 2001, 2003).For example, Tarone and Kuehn show that even native-speaker applicants must


learn that the interviewer does not want to hear the story of why an applicant isapplying for assistance; instead, the interviewer needs to have other informationto make the determination of eligibility. In this example from Tarone and Kuehn(2000, p. 112), the interviewer cuts off the applicant's explanation of why he isapplying for assistance and moves to the next stage of the intake interview, theexplanation of an applicant's rights and responsibilities.

(1) Social services interview

Applicant: Right, I've been having a problem trying to get it, and so-Interviewer: OK. We'll go over your rights and responsibilities

Examples of such teaching are found throughout the reports on institutionaltalk. In the exchange between the teacher and the student reported by Churchill(1999, p. 16), the teacher indicates that he expects a student to say somethingwhen turning in an assignment, and the student responds by constructing arequest with the help of her friend.

(2) Mariko attempts to hand in her homework assignment to her teacher

Teacher: What? You have nothing to say?Mariko: Could (looks at Yukiko)Teacher: Could you...Mariko: Could you check my homework.

Advisors in academic advising interviews also provide opportunities forgraduate students to take an appropriate turn by providing competency slots.Advisors open the core of the advising session (the "directive" phase, Agar,1985) with questions that indicate to the student-participants that they areexpected to have suggestions ready; and through these questions, the studentsmay learn that they should perform such a speech act in this context, as in thefollowing examples from Bardovi-Harlig & Hartford (1996, p. 175):

(3) Competency slots

a. Now, what do you want to take in the fall?b. Okay....so you looked through the list of courses, so you pretty much know

what you want to take?c. Do you have some idea of what you would want to take?

At the same time, students are given specific opportunities by the advisor todisplay their competence as graduate students through institutional talk. Anadvisor may also express his approval of turns taken by a student as the advisordid in Example (4) (Bardovi-Harlig & Hartford, 1996, p. 176). This may alsoserve to indicate to the student that he or she is learning the rules of the genre.


(4) Explicit acknowledgment of the rules of the genre

Advisor: I wish I had people come in with, that knew what, that knew whatthey wanted to do, like you. That helps me.

These exchanges are examples of the input available to institutionalparticipants to help them learn the rules of institutional talk. Transcriptions ofemployment interviews and the accompanying retrospective interviews thatreveal unresolved miscommunication stemming from either lack of knowledgeor conflicting beliefs about employment interviews (Kerekes, 2001, 2003),suggest that some types of institutional talk, or some institutional representa-tives, might be less likely to assist novices in learning the genre. We do not yetknow what type of input is particularly effective or how learners (i.e., noviceclients or members, whether language learners or not) might interpret certainrules for their own purposes (for example, Tarone & Kuehn, 2000, report thateven experienced applicants try to tell the reasons for their application), but thestudy of novices engaged in institutional talk will lead to answers of theseimportant questions and perhaps to important insights regarding learning andinput more generally.

Institutional talk provides a microcosm of culture, and observing noviceparticipants as they become experts allows researchers to observe how institu-tional talk and institutional rules are taught and learned. Because institutionsform closed and definable systems (or relatively more easily definable systemscompared to the whole of culture) and are culture specific, researching thelearning of institutional rules through institutional talk breaks the study ofcultural rule learning into manageable proportions.

ADDITIONAL CONSIDERATIONS

Generalizability

The features that ensure a high level of comparability are also those features thatmight restrict a broader interpretation of studies of institutional talk to otherdiscourses. In institutional talk, as we have shown, participants have fixed rolesand fixed goals, and while this promotes comparability across interactions,findings may not apply to other settings. In institutional talk participantsgenerally have a single institutional status within an individual encounter, andeven within a series of encounters. For example, the advisor is always of higherinstitutional status than the student within the academic advising session. Inaddition, Ellis (1992) and Churchill (1999) note that students talk to a limitedrange of addressees, namely the teacher and other classmates. Participants havegenerally fixed roles, as determined by the nature of the institutional contextitself: For example, within a single psychiatric interview, the nurse does theassessment, and the patient responds to her questions; within the social servicesinterview, the financial worker asks questions to verify the financial status of the


applicant and the applicant supplies the information. We do not generally havethe opportunity to observe the interviewers switch roles and become therespondent, as we might in casual conversations. However, following a singleinstitutional participant through multiple types of interaction balances the datafor the effect of fixed roles in a single type of interaction. For example, Cameronand Williams's (1997) study of talk in a hospital psychiatric unit shows onespeaker in two institutional roles each with a different status. The main speaker,an NNS nurse, is the institutional representative and higher status speaker whenshe interviews her NS patients, and is a nurse-trainee of lower status when shediscusses her case load with the nursing supervisor, dyne's (1994) corpus oftalk among institutional members includes key participants talking to bothsupervisors and peers. Thus, as the work by Clyne (1994) and Cameron andWilliams (1997) shows, solutions to some of the limitations in design can befound in the institutional setting itself.

There is no doubt that institutional talk is different from casual conversationin many ways, as outlined earlier. In casual conversations, establishing comity,or friendly relations, is key (Ashton, 1993). Therefore, what we learn from thestudy of institutional talk should only be applied with great caution toconversation more generally. At the same time, however, observations of institu-tional talk could serve to generate research hypotheses that can be tested againstcasual conversation in interlanguage pragmatics research.

Discourse within individual institutions may be generalizable within theinstitutions themselves. It is important to note that the study of institutional talkfor its own sake in interlanguage pragmatics is not trivial. Institutional encoun-ters occur fairly frequently in the lives of people and should not be regarded asrare or unusual types of speech events. The number of normative speakers atinstitutions in the anglophone world is significant and growing. At our researchuniversity alone, in the 2002-2003 academic year, there were more than 3,320,or 8.5% international students, and the expectation is that that number will growto one student in ten in coming years. The number of international students atseveral major research universities in California and New York exceeded 4,500,eclipsing the 10% figure (Opendoors, 2001). The number of internationalstudents studying in the United States during the same period was 514,723(Opendoors, 2001). This is not an incidental population. With increasingnumbers of normative speakers engaged in institutions in the areas of education,health care, social services, and the workplace, institutional talk becomes asignificant area of investigation in its own right. Although the number ofinternational students in the United States and other anglophone countries issignificant, it does not even begin to approach the number of immigrants inthese (and other) countries who most certainly will have numerous institutionalencounters.

Finally, although it may seem that institutional discourse represents only avery narrow set of linguistic and pragmatic activities in which native andnonnative speakers might engage, in fact, many of the activities may berepresentative of a more general type of discourse, such as seeking help and


solving problems, so that the practice of such activities could be carried beyondthese settings by language learners (as well as native speakers).

Ethical Issues

The last methodological consideration is also an ethical one. Most researchinstitutions require that the researchers obtain approval from the appropriatecommittee to conduct research with human subjects. In addition, the institutionswhich are the objects of research may also have some concerns. Often the talkwhich is targeted for research is regarded as "private" or in some other way tobe protected from public view. For the researcher, in practical terms, this meansgetting consent from all of the participants, some of whom will be bound bycertain institutional concerns. Furthermore, as Sarangi and Roberts (1999, pp.41-42) point out, there is concern among researchers about whether or not suchresearch should have tangible benefits for the people being studied. Thisincludes the questions of how research findings should be disseminated: Is theaudience other academics and research communities, with no access availablefor the institutional participants?9 It seems to us that in the field of appliedlinguistics, these concerns of application and dissemination may be less trouble-some than for "pure" researchers. That is, one of the purposes of appliedlinguistics dissemination is to educate language teachers about the discourses ofsuch institutions, where, in turn, they will educate their learners who arepotential participants, both as clients and workers. (See, for example, Li, 2000;Tarone & Kuehn, 2000). It is less obvious, perhaps, how the institutions willdirectly benefit, and here it would be up to the researcher to make that clear tothe institution in question.

Although such concerns must be taken into consideration by the investiga-tors, we should note that it is possible to conduct such research in hospital ormedical settings, even though concerns for privacy might seem especiallyrelevant and perhaps major obstacles. The study by Cameron and Williams(1997), for example, was a result of the second author participating in alanguage needs analysis for normative speaking graduate nursing students, sothat the application, or benefit, "preceded" the later analysis (p. 423). In afootnote, however, these authors discuss a number of the problems related toresearch on institutional discourse and medical discourse, in particular (pp. 439-440, fn 5). These difficulties include real-world concerns such as opening theinstitution to legal actions such as being sued, the possibility of the institutionwithdrawing support at any time, limited access to follow-up interviews withparticipants, and numerous other considerations. Tarone and Kuehn (2000) alsodiscuss the issues of privacy in the social service interview. In contrast, studiesby Bardovi-Harlig and Hartford (1990, 1993a), Bouton (1995), Churchill(1999), Thonus (1998, 1999), Tyler (1995) and Tyler & Davies (1990), while

9 By "access" we mean that the academic field may employ a discourse that is so particular to thefield that nonprofessionals are not able to easily engage it.


maintaining the anonymity of their informants, point to the advantage ofsecuring permission to study discourse at one's own institution.

Nevertheless, institutions and the people who make up institutions often doagree to be studied, and in a remarkable array of settings. In addition to thesettings described in Tables 1.2 and 1.3, other institutional settings include thosecited earlier this chapter, some of which can be found in other chapters of thisbook. Any serious researcher can approach an institution. It need not be onlyresearchers with years of experience. Among the studies listed in Tables 1.2 and1.3, a number were carried out by doctoral candidates for their doctoral theses.We discuss in some detail the practical considerations of conducting institutionalresearch in the final chapter of this volume.

CONCLUSION

In this chapter we have addressed a primary concern in research on interlan-guage pragmatics: the pull between replicable, comparable data and authentic,situated data. We have argued that one promising avenue of compromise,institutional discourse, yields data that are both comparable and authentic, andwhich meet the three primary requirements of comparability, interactivity, andconsequentiality. Institutional discourse is comparable because institutionalspeech events contain similar attributes within institutions and sometimes acrossthem: They tend to have similar time-frames, similar structuring, similar topicalselection, and similar participant roles. They are interactive because they includeat least two participants who exchange talk for the institutional purposes forwhich they meet. Finally, they are consequential because they are situated inauthentic situations and have goals and outcomes of concern to all theinteractants.

We have also demonstrated that institutional discourse may be utilized ininterlanguage pragmatics research not only for the investigation of a singleaspect of pragmatics, speech acts, but for other aspects as well, includingconversational structuring, turn contributions, placement and timing of contribu-tions, and politeness strategies.

Institutional discourse can be utilized to investigate the acquisition ofinterlanguage pragmatics as well as instances of usage. We have shown thatoften both native and nonnative speakers have to learn such discourses, andargue that this is an excellent opportunity for the researcher to compare acquisi-tional strategies across native and nonnative speakers.

Institutional discourse can also be investigated for its own sake withininterlanguage pragmatics because of the number of nonnative speakers who maybe involved in such events. We have argued that the information gained from theanalysis of institutional discourse has practical application. It is a type ofdiscourse that almost all second language learners will have to engage in in thereal world at some point in their learning histories. Pedagogical usages can bedeveloped which will be meaningful to the learners' lives. Although not alllearners will encounter all institutional events, there are some which may be


seen as occurring and recurring for them and for which they can acquire someexpertise.

Finally, although the investigation of institutional talk is not a substitute forthe examination of casual conversation which interlanguage pragmatics studieswill eventually need to undertake, it can lead to fruitful investigation on anumber of planes for interlanguage pragmatics research. Furthermore, it willhelp bring interlanguage pragmatics research in line with work already beingcarried out by other discourse analysts. Clearly, the amount of research focusingon the discourse of institutional talk which is done from frameworks other thaninterlanguage pragmatics suggests that it is an important and promising area forsuch research. There is every reason to believe that the same holds true forinterlanguage pragmatics research.

REFERENCES

Akinnaso, F. N., &. Ajirotutun, C. (1982). Performance and ethnic style in job interviews. In J.Gumperz (Ed.), Language and social identity (pp. 145-62). Cambridge, England: CambridgeUniversity Press.

Agar, M. (1985). Institutional discourse. Text, 5, 147-168.Aston, G. (1993). Notes on interlanguage comity. In G. Kasper & S. Blum-Kulka (Eds.),

Interlanguage pragmatics (pp. 224-250). Oxford, England: Oxford University Press.Bardovi-Harlig, K. (1999a). Researching method. In L. F. Bouton (Ed.), Pragmatics and language

learning (Vol. 9, pp. 237-264). University of Illinois, Urbana-Champaign: Division of Englishas an International Language.

Bardovi-Harlig, K. (1999b.) The interlanguage of interlanguage pragmatics: A research agenda foracquisitional pragmatics. Language Learning, 49, 677-713.

Bardovi-Harlig, K. (1996). Pragmatics and language teaching: Bringing pragmatics and pedagogytogether. In L. F. Bouton (Ed.), Pragmatics and language learning (Vol. 7, pp. 21-39).University of Illinois, Urbana-Champaign: Division of English as an International Language.

Bardovi-Harlig, K., & Dornyei, Z. (1998). Do language learners recognize pragmatic violations?Pragmatic vs. grammatical awareness in instructed L2 learning. TESOL Quarterly, 32, 233-259.

Bardovi-Harlig, K., & Hartford, B. S. (1990). Congruence in native and normative conversations:Status balance in the academic advising session. Language Learning, 40,467-501.

Bardovi-Harlig, K., & Hartford, B. S. (1991). Saying "No": Native and normative rejections inEnglish. In L. F. Bouton & Y. Kachra (Eds.), Pragmatics and language learning (Vol. 2, pp.41-57). University of Illinois, Urbana-Champaign: Division of English as an InternationalLanguage.

Bardovi-Harlig, K., & Hartford, B. S. (1993a). Learning the rules of academic talk: A longitudinalstudy of pragmatic development. Studies in Second Language Acquisition, 15, 279-304.

Bardovi-Harlig, K., & Hartford, B. S. (1993b). Refining the DCT: Comparing open questionnairesand dialogue completion tasks. In L. F. Bouton & Y. Kachru (Eds.), Pragmatics and languagelearning (Vol. 4, pp. 143-165). University of Illinois, Urbana-Champaign: Division of Englishas an International Language.

Bardovi-Harlig, K., & Hartford, B. S. (1993c). The language of comembership. Research onLanguage and Social Interaction, 26, 227-257.

Bardovi-Harlig, K., & Hartford, B. S. (1996). Input in an institutional setting. Studies in SecondLanguage Acquisition, 18, 171-188.

Bardovi-Harlig, K., & Hartford, B. S. (1997, April). Academic Advising Sessions with Students in anIntensive English Program. Paper presented at the Eleventh International Conference onPragmatics and Language Learning, Urbana, IL.


Beebe, L. (1994, March). Notebook data on power and the power of notebook data. Paper presentedat the Twenty-eighth Annual TESOL Conference, Baltimore, MD.

Beebe, L., & Cummings, M. C. (1985). Speech act performance: A function of the data collectionprocedure? Paper presented at the Nineteenth Annual TESOL Conference, New York, NY.

Beebe, L., & Cummings, M. C. (1996). Natural speech act data versus written questionnaire data:How data collection method affects speech act performance. In S. M. Gass & J. Neu (Eds.),Speech acts across cultures: Challenges to communication in a second language (pp. 65-86).Berlin: de Gruyter.

Bergman, M. L., & Kasper, G. (1993). Perception and performance in native and nonnative apology.In G. Kasper & S. Blum-Kulka (Eds.), Interlanguage pragmatics (pp. 82-107). Oxford,England: Oxford University Press.

Blum-Kulka, S., & Sheffer, H. (1993). The metapragmatic discourse of American-Israeli families atdinner. In G. Kasper & S. Blum-Kulka (Eds.), Interlanguage pragmatics (pp. 196-223).Oxford, England: Oxford University Press.

Bouton, L. F. (1995). A cross-cultural analysis of the structure and content of letters of reference.Studies in Second Language Acquisition, 17, 211-244.

Button, G. (1987). Moving out of closings. In G. Button & J. R. E. Lee (Eds.), Talk and socialorganization (pp. 101-151). Clevedon, England: Multilingual Matters.

Cameron, R., & Williams, J. (1997). Sentence to ten cents: A case study of relevance andcommunicative success in nonnative-native speaker interaction in a medical setting. AppliedLinguistics, 18, 415-445.

Churchill, E. F. (1999, October). Pragmatic development in L2 request strategies by lower levellearners. Paper presented at the Second Language Research Forum, Minneapolis, MN.

Clyne, M. (1994). Inter-cultural communication at work: Cultural values in discourse. Cambridge,England: Cambridge University Press.

Cohen, A. D. (1996). Speech acts. In S. L. McKay & N. H. Hornberger (Eds.), Sociolinguistics andlanguage teaching (pp.383-420). Cambridge, England: Cambridge University Press.

Cohen, A. D., & Olshtain, E. (1993). The production of speech acts by EFL learners. TESOLQuarterly, 27, 33-56.

Cohen, A. D., & Olshtain, E. (1994). Researching the production of second language speech acts. InE. E. Tarone, S. M. Gass, & A. D. Cohen (Eds.), Research methodology in second-languageacquisition (pp. 143-156). Hillsdale, NJ: Lawrence Erlbaum Associates.

Conley, J., & OBarr, W. (1990). Rules versus relationships: The ethnography of legal discourse.Chicago: University of Chicago Press.

Cook-Gumperz, J., & Messerman, L. (1999). Local identities and institutional practices: constructingthe record of professional collaboration. In S. Sarangi & C. Roberts (Eds.), Talk, work andinstitutional order: Discourse in medical, mediation and management settings (pp. 145-182).Berlin: de Gruyter.

Douglas, D., & Selinker, L. (1994). Research methodology in context-based second-languageresearch. In E. E. Tarone, S. M. Gass, & A. D. Cohen (Eds.), Research methodology in second-language acquisition (pp. 119-131). Hillsdale, NJ: Lawrence Erlbaum Associates.

Drew, P., & Heritage, J. (1992). (Eds.) Talk at work. Cambridge, England: Cambridge UniversityPress.

Ellis, R. (1992). Learning to communicate in the classroom: A study of two learners requests.Studies in Second Language Acquisition, 14, 1-23.

Erickson, F. (1999). Appropriation of voice and presentation of self as a fellow physician: Aspects ofa discourse of apprenticeship in medicine. In S. Sarangi & C. Roberts (Eds.), Talk, work andinstitutional order: Discourse in medical, mediation and management settings (pp. 109-144).Berlin: de Gruyter.

Erickson, F., & Shultz, J. (1982). The counselor as gatekeeper. New York: Academic Press.Fisher, S., & Todd, A. (Eds.). (1983). The social organization of doctor-patient communication.

Norwood, N.J.: Ablex.Garton, M. (2000). The effect of age, gender, and degree of imposition on the production of native

speaker and nonnative speaker requests in Hungarian. Unpublished doctoral thesis, IndianaUniversity, Bloomington, Indiana.


Gass, S. M., & Neu, J. (1996). (Eds.). Speech Acts across cultures: Challenges to Communication ina Second Language. Berlin: de Gruyter

Gass, S. M., & Houck, N. (1999). Interlanguage Refusals. Berlin: de GruyterGass, S. M. & Varonis, E. M. (1985).Variation in native speaker speech modification to normative

speakers. Studies in Second Language Acquisition, 7, 37-57.Geis, M. L., & Harlow, L. L. (1996). Politeness strategies in French and English. In S. M. Gass & J.

Neu (Eds.), Speech acts across cultures: Challenges to communication in a second language(pp. 129-153). Berlin: de Gruyter.

Goffman, E. (1971). Relations in public. New York: Basic Books.Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Speech acts (Syntax and

Semantics, Vol. 3, pp. 41-58). New York: Academic Press.Gumperz, J. (1982a). Fact and inference in courtroom testimony. In J. Gumperz (Ed.), Language and

social identity (pp. 163-194). Cambridge, England: Cambridge University Press.Gumperz, J. (1982b). Language and social identity. Cambridge, England: Cambridge University

Press.Hartford, B. S., & Bardovi-Harlig, K. (1992). Closing the conversation: Evidence from the academic

advising session. Discourse Processes, 15, 93-116.Hartford, B. S., & Bardovi-Harlig, K. (1996a). "At your earliest convenience": Written student

requests to faculty. In L. F. Bouton (Ed.), Pragmatics and language learning, (Vol. 7, pp. 55-69). University of Illinois, Urbana-Champaign: Division of English as an InternationalLanguage.

Hartford, B. S., & Bardovi-Harlig, K. (1996b, August). Cross-sectional study of the acquisition ofspeech acts by non-native learners of English: The use of natural data in an intensive Englishprogram. Paper presented at AILA, Finland.

He, A. W. (1994). Withholding academic advice: Institutional context and discourse practice.Discourse Processes, 18, 297-316.

Houck, N., & Gass, S. (1996). Non-native refusals: A methodological perspective. In S. M. Gass &J. Neu (Eds.), Speech acts across cultures: Challenge to communication in a second language(pp. 45-64). Berlin: de Gruyter.

Hymes, D. (1964). Models of interaction of language and social life. In D. Hymes & J. Gumperz(Eds.), Directions in Sociolinguistics: The Ethnography of Communication (pp. 35-71). NewYork: Holt, Rinehart, and Winston.

Irvine, J. (1974). Strategies of status manipulation in Wolof greetings. In R. Bauman & J. Scherzer(Eds.), Explorations in the ethnography of speaking (pp. 167-191). Cambridge, England:Cambridge University Press.

Johnston, B., Kasper, G., & Ross, S. (1994). Effect of rejoinders in production questionnaires.University of Hawaii Working Papers in ESL, 13, 121-143.

Johnston, B., Kasper, G., & Ross, S. (1998). Effect of rejoinders in production questionnaires.Applied Linguistics, 19, 157-182.

Kasper, G. (1991, October). (More on) Methods of data collection in interlanguage pragmatics.Paper presented at the Conference on Theory Construction and Methodology in SecondLanguage Acquisition Research. East Lansing, MI.

Kasper, G. (1997). The role of pragmatics in language teacher education. In K. Bardovi-Harlig & B.S. Hartford, (Eds.), Beyond Methods: Components of language teacher education (pp. 113-136). New York: McGraw Hill.

Kasper, G. (1998). Interlanguage pragmatics. In H. Byrnes (Ed.), Learning foreign and secondlanguage: Perspectives in research and scholarship (pp. 183-208). New York: The ModernLanguage Association.

Kasper, G., & Blum-Kulka, S. (1993). (Eds.). Interlanguage pragmatics. Oxford, England: OxfordUniversity Press.

Kasper, G., & Dahl, M. (1991). Research methods in interlanguage pragmatics. Studies in SecondLanguage Acquisition, 13, 215-247.

Kasper, G., & Schmidt, R. (1996). Developmental issues in interlanguage pragmatics. Studies inSecond Language Acquisition, 18, 149-169.

Kerekes, J. (2001) The co-construction of a successful gatekeeping encounter: Strategies of linguisti-cally diverse speakers. Unpublished doctoral thesis, Stanford University, Stanford, California.


Kerekes, J. (2003). Distrust: A determining factor in the outcomes of gatekeeping encounters. In S.Ross (Ed.), Misunderstanding in spoken discourse (pp. 227-257). London: Longman/Pearson.

Kress, G., & Fowler, R. (1979). Interviews. In R. Fowler, B. Hodge, G. Kress, & T. Trew (Eds.),Language and control (pp. 63-80). London: Routledge and Kegan Paul.

Kuha, M. (1997). The computer-assisted Interactive DCT: A study in pragmatics research methodol-ogy. In L. Bouton (Ed.) Pragmatics and language learning, Monograph 8 (pp. 99-127).University of Illinois, Urbana-Champaign: Division of English as an International Language.

Labov, W., & Fanshel, D. (1977). Therapeutic Discourse. New York: Academic Press.Leary, A. (1994). Conversational replicas on computer for intermediate-level Russian: A Theoreti-

cal Model of a Design Paradigm. Unpublished masters thesis, University of Iowa, Iowa City.Levinson, S. (1992). Activity types and language. In P. Drew., & J. Heritage. (Eds.), Talk at work

(pp. 66-100). Cambridge, England: Cambridge University Press,Li, D. (2000). The pragmatics of making requests in the L2 workplace: A case study of language

socialization. Canadian Modern Language Review, 57, 58-87.Linde, C. (1988). The quantitative study of communicative success: Politeness and accidents in

aviation discourse. Language in Society, 17, 375-400.Mehan, H. (1994). The role of discourse in learning, schooling, and reform. In B. McLeod (Ed.),

Language and learning: Educating linguistically diverse students (pp. 71-94). Albany, NY:State University of New York Press.

Moran, M. H., & Moran, M. G. (1985). Business letters, memoranda, and resumes. In M.G. Moran& C. Journet (Eds.) Research in technical communication (pp. 313-52). Westport, CT:Greenwood Press.

Murphy, B., & Neu, J. (1996). My grade's too low: The speech act set of complaining. In S. M.Gass & J. Neu (Eds.), Speech acts across cultures: Challenges to communication in a secondlanguage (pp. 191-216). Berlin: de Gruyter

ODonell, K. (1990). Difference and dominance: How labor and management talk conflict. In A.Grimshaw (Ed.), Conflict talk (pp. 210-40), Cambridge, England: Cambridge University Press.

Omar, A. S. (1992). Opening and closing conversations in Kiswahili: A study of the performance ofnative speakers and learners. Unpublished doctoral thesis, Indiana University, Bloomington,Indiana.

Phillips, S. (1990). The judge as third party in American trial-court conflict talk. In A. Grimshaw(Ed.), Conflict talk (pp. 197-209), Cambridge, England: Cambridge University Press.

Rodriquez, S. (2001). The perception of requests in Spanish by instructed learners of Spanish in thesecond- and foreign-language contexts: A longitudinal study of acquisitional patterns.Unpublished doctoral thesis, Indiana University, Bloomington, Indiana.

Rose, K. (1997). Film in interlanguage pragmatics research. Perspectives (Working Papers of theDepartment of English, City University of Hong Kong), 9, 111-144.

Rose, K. (2000). An exploratory cross-sectional study of interlanguage pragmatic development.Studies in Second Language Acquisition, 22, 27-67.

Rose, K., & Ono, R. (1995). Eliciting speech act data in Japanese: The effect of questionnaire type.Language Learning, 45, 191-223

Sarangi, S., & Roberts, C. (Eds.). (1999). Talk, work and institutional order: Discourse in medical,mediation and management settings. Berlin: de Gruyter.

Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, VIII, 289-327.Schmidt, R., Shimura, A., Wang, Z., & Jeong, H. (1996). Suggestions to buy: Television

commercials from the U.S., Japan, China, and Korea. In S. M. Gass & J. Neu (Eds.), Speechacts across cultures: Challenges to communication in a second language (pp. 285-316).Berlin: de Gruyter

Selinker, L., & Douglas, D. (1985). Wrestling with 'context' in interlanguage theory. AppliedLinguistics, 6, 75-86.

Selinker, L., & Douglas, D. (1988). Using discourse domains in creating interlanguage: Context,theory, and research methodology. In J. Klegraf & D. Nehls (Eds.), Essays on the Englishlanguage and applied linguistics on the occasion of Gerhard Nickels 60th birthday (pp. 357-379). Heidelberg: Julius Groos.

Selinker, L., & Douglas, D. (1989). Research methodology in context-based second-languageresearch. Second Language Research, 5, 93-126.

3 6 Bardovi-Harlig and Hartford

Tannen, D. (1986). Conversational style: Analyzing talk among friends. Norwood, NJ: Ablex.Tannen, D., & Wallat, C. (1993). Interactive frames and knowledge schemas in interaction:

Examples from a medical examination/interview. In D. Tannen (Ed.), Framing in discourse(pp. 57-78). Oxford, England: Oxford University Press.

Tarone, E., & Kuehn, K. (2000). Negotiating the social service oral intake interview: Communica-tive needs of normative speakers of English. TESOL Quarterly, 34, 99-126.

Thonus, T. (1998). What makes a writing center tutorial successful: An analysis of linguisticvariables and social context. Unpublished doctoral thesis, Indiana University, Bloomington,Indiana.

Thonus, T. (1999). How to communicate politely and be a tutor, too: NS-NNS interaction andwriting center practice. Text, 19, 253-279.

Thonus, T. (2002). Tutor and student assessments of academic writing materials: What is "success?"Assessing Writing, 8, 110-134.

Tyler, A. (1995). The coconstruction of cross-cultural miscommunication: Conflicts in perception,negotiation and enactment of participant role and status. Studies in Second LanguageAcquisition, 17, 129-152.

Tyler, A., & Davies, C. (1990). Cross-linguistics communication missteps. Text, 10, 385-411.Woken, M. D., & Swales, J. M. (1989). Expertise and authority in native-normative conversations:

The need for a variable account. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.),Variation in second language acquisition (Discourse and Pragmatics, Vol. 1, pp. 211-222).Clevedon, England: Multilingual Matters.

Wolfson, N. (1986). Research methodology and the question of validity. TESOL Quarterly, 20, 689-699.

Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. Cambridge, MA: Newbury House.Yates, L. (2000, July). "Ciao, Guys!": Mitigation addressing positive and negative face concerns in

the directives of native-speaker and Chinese background speakers of Australian English.Unpublished doctoral thesis, LaTrobe University, Bundoora, Victoria, Australia.

Yli-Jokipii, H. (1991). Running against time and technology: Problems in empirical research intowritten business communication. In K. Sajavaara, D. March, & T. Keto (Eds.), Communicationand discourse across cultures (pp. 59-72). Jyvaskyla, Finland: Kopi-Jyva Oy.

Yuan, Y. (1998). Sociolinguistic dimensions of the compliment speech event in the southwestMandarin dialect spoken in Kunming, China. Unpublished doctoral thesis, Indiana University,Bloomington, Indiana.

Zuengler, J. (1989). Performance variation in NS-NNS interactions: Ethnolinguistic difference, ordiscourse domain? In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation insecond language acquisition (Discourse and Pragmatics, Vol. 1, pp. 228-244). Clevedon,England: Multilingual Matters.

2

Writing Center Interaction:Institutional Discourse and the Role of Peer

Tutors

Jessica WilliamsUniversity of Illinois

There are many different interactional settings in a university, including large,teacher-fronted classes, seminars, advising sessions, and student—teacher confer-ences, and new students must learn the pragmatic rules for each of thesesettings. They will interact with fellow students, teaching assistants, tutors,professors, and administrators, and with each, expectations and norms forinteraction may differ. For second language (L2) learners, this poses a particularchallenge, with high language proficiency no guarantee of pragmatic success(Bardovi-Harlig, 2001; Bardovi-Harlig & Hartford, 1990, 1993b; Fiksdal, 1989;Harris, 1997; He, 1993; Kasper & Rose, 1999; Storch, 2002; Thonus, 1999a,1999b; Tyler, 1995). One setting that L2 learners may encounter with increasingfrequency is the university writing center. The focus of this study is theinteraction between tutors and writers seeking their assistance. Specifically, itaddresses the institutional nature of this interaction, the status of tutors withinthose sessions, and whether they behave differently when they work with L2writers versus native English (NE) writers1.

Writing centers or labs have been established at many if not most majoruniversities in the country to provide assistance and support for writers acrossthe institution. There are a variety of models for writing centers (Kinkead &Harris, 1993), but one common model uses peer tutoring, that is, the tutors whoprovide assistance are themselves students. Most writing centers that operate inthe peer-tutoring tradition trace their roots to the work of Bruffee (1984), whomaintained that tutoring is essentially an interaction between peers who sharesimilar backgrounds, experience, and status, one that creates a different andpowerful context for learning. Trimbur (1987), in his discussion of the apparentcontradiction of being both a peer and a tutor, sees the writing center conferenceas a colearning situation, where the tutor and writer collaborate and negotiatemeans and goals as equals. The value of peer tutoring is that it allows unfetteredconversation. The peers are not miniteachers; rather, they represent a differentset of eyes and ears, another voice in the writing process. They are ofteninterpreters who have one foot in the teacher's discourse community and another

1 The term NE writer is used here since the normal term, native speaker, is not a suitable description.Some L2 writers are indistinguishable from native speakers in their oral production.

37

38 Williams

in the students' (Harris, 1995). Importantly, these peers have recently gonethrough similar experiences, are closer to the writers' experiences, and thereforemay better understand the challenges they face. The literature on writing centersin general, and peer tutoring in particular, is extensive (e.g., Bouquet, 1999;Gillespie, Gillam, Brown, & Stay, 2002; Kinkead & Harris, 1993; Murphy &Law, 1995; two dedicated journals: Writing Lab Newsletter and Writing CenterJournal), yet there has been relatively little investigation of L2 writers'experiences in the writing center, specifically, the interaction between L2 writersand their tutors (though see Blau, Hall, & Strauss, 1998; Ritter, 2002; Thonus1999a, 1999b, 2001, 2002). Much of the writing on L2 writers in the writingcenter is limited to a cautious set of do's and don'ts (e.g., Gadbow, 1992; Harris& Silva, 1993; Kennedy, 1993; Moser, 1993; Powers, 1993; for a somewhatdifferent perspective, see Blau & Hall, 2002).

The role of the peer tutor can be ambiguous and delicate, combining as itdoes a status that is equal yet somehow unequal to the writer-client. Mostliterature on peer tutoring, at least among children, assumes that that the twoparticipants "do not have equal status in their instructional relationship. Theirengagement, therefore, is low on equality, at least relative to other forms of peerdiscourse" (Damon & Phelps, 1989, p. 11). The notion of peerness, howevervexed it may be when applied to tutors of native speakers, becomes even moreproblematic in interaction with L2 learner. It is an open question whether nativespeaking tutors, who bear the inherent authority of native speaker status, and L2writers can really be said to operate on an equal footing, a question that will beexplored in this study.

WRITING CENTER INTERACTION AS INSTITUTIONAL DISCOURSE

"Conversation" is a popular word in academia, composition studies, and writingcenters. The handbook for the University of Illinois at Chicago's Writing Centeropens with the following statement, "Our goal is to help each other conductmeaningful conversations with students who come to us for help with writing."(Aleksa, Bednarowicz, Smith, Brecke, & Huang, 2000, p. 5, emphasis added).In fact though, writing center sessions are not really conversations in theeveryday sense; rather, they are more characteristic of institutional discourse.The literature on institutional discourse distinguishes between classic institu-tional discourse (between institutional representatives and service seekers) andworkplace talk (between institutional workers) (Bardovi-Harlig & Hartford,chapter 1, this volume). Interaction between writers and tutors falls into the firstcategory. Within this category, we can again distinguish between two kinds ofpairings: The first consists of experts providing some sort of service to a clientseeking this service. This is typical, for instance, of interaction in medicalsettings (e.g., Fisher & Todd, 1983; Cameron & Williams, 1997; Tannen &Wallat, 1993) and gatekeeping encounters (e.g., Erickson & Schultz, 1982; He,1994; Kerekes, chapter 4, this volume; Tarone & Kuehn, 2000). Second is thepairing between expert and novice, more frequent in educational or training

Writing Center Interaction 39

settings, perhaps between teacher and student or apprentice (e.g., Lave &Wenger, 1991; Mehan, 1985; Rudolph, 1994; Woken & Swales, 1989). Thecurrent study is of interest for two reasons. First, writing center sessions stand atthe intersection of these two types of interaction: Writers come to the centerseeking services, as they might in a medical encounter, yet they are alsopotentially in the position of novice or apprentice academic writer. Second, therehas been relatively little research done on this kind of institutional discourseinvolving L2 speakers (see Bardovi-Harlig & Hartford, chapter 1, this volume,for a review). The study has three parts: It begins with an examination of thestructure of writing center interaction from the perspective of institutionaldiscourse, then addresses how tutor status informs and is informed by this inter-acttion, and finally, explores the extent to which the interaction between tutorsand L2 and NE writers is comparable.

Data Collection

The corpus for this study consists of 10 writing center sessions. There are fourtutor participants, all fluent speakers of English. One is a monolingual nativespeaker, two are bilingual native speakers (L2s = Spanish, Tagalog), and one is ahighly fluent, bilingual L2 speaker of English (L1 = Russian). The writerparticipants include two NE writers, both of whom turned out to be bilingual aswell (L2s = Korean, Greek), though neither their oral nor spoken productionshowed signs of the nonEnglish language, and three L2 writers (L1s = Chinese(2), Khmer). This kind of ethnic and linguistic diversity is the rule rather thanthe exception on this urban, largely commuter campus, where white mono-lingual English speakers are in the minority. Furthermore, white middle-class,monolingual speakers with good high school preparation often test out of thecomposition classes from which the participants were drawn. It is therefore notsurprising that even the "NS baseline" group as well as the tutors present amulticultural, multilingual profile. Each writer participated in two sessions, inall but one case, with different tutors. Both tutors and writers were paid for theirparticipation in the study. Writers were recruited from the basic English compo-sition classes required of all undergraduates who place into it. Tutors wererecruited from the regular staff at the University Writing Center.

The corpus brings together several types of data in an effort to provide acomplete view of the session as well as the real life consequences of thesessions: background interviews, tutoring sessions, and stimulated recalls. Allparticipants were initially interviewed about their backgrounds, their expec-tations as writers or tutors, and their past experience tutoring L2 writers/usingthe writing center. All tutoring sessions were videotaped, transcribed, and codedby the researcher and a research assistant. Finally, the corpus includes amodified stimulated recall of the session with the writer and the tutor. Eachparticipant was interviewed separately on audiotape within 3 days of eachsession while watching the videotape (Gass & Mackey, 2000). This final step inthe triangulation of the data was included in order to analyze the participants'

40 Williams

motivation for their contributions to the interaction and their understanding of,and reactions, to their partners' contributions. The combination of the videotapesand recall sessions helped to determine, first with quantitative measures, andthen through participant reflection, the roles that the participants played and howthey viewed the role of their partner in the session. In addition to the transcribedvideotapes, the preliminary drafts or predraft work that the writer brought to thesession were copied and collected. The writers were also asked to submit a copyof any subsequent draft that emanated from work during the session. Analysis ofthe written data is not presented in this article (see Williams, in press).

THE STRUCTURE OF INTERACTION

One of the identifying characteristics of institutional discourse is its predictablestructure (Agar 1985; Drew & Heritage, 1992; Sarangi & Roberts, 1999). Thereare a number of reasons for this, the primary one being its task or goal-orien-tation. The topics that can be discussed are narrowly constrained and the roles ofthe participants, though subject to negotiation, are usually clearly drawn andgenerally of unequal status. Agar maintains that the goal orientation and specificprocedures associated with institutional discourse result in a predictablesequence of moves. He proposes that institutional discourse generally has threephases, diagnosis, directive, and report. These are widely reported in medicaland management settings, but also in advising and tutoring sessions (Bardovi-Harlig & Hartford, 1993a; Ritter, 2002; Thonus, 1999a). Even before thediagnosis phase, however, in writing center sessions, there is often an overtgoal-setting phase that may be absent from other institutional interaction. Afterintroductions, tutors will elicit the writer's goals for the session. They normallywill also extend an offer of help, which contains the implicit claim to theirability to provide that help, that is, their expertise. In example (1) the writer, aNE writer, provides a statement of her problem, and an admission of her novicestatus.

(1) T = tutor (Joe) W = writer (Esther)2

T: OK. My name is Joe3.W: My name is Esther (shake hands)T: How can I help you today?W: Well, I have to write.like.a compare and contrast paper on religion. How

two authors have wrote.and I have to write a paper about.what I thinkum.whether or not we need religion....um. But I'm kind of stuck. / don'tknow really how to write a compare and contrast paper. I need some helpwith that.

T: Do you have um.the assignment sheet that your teacher handed out?W: Yes. This is uh.the assignment sheet. She emailed us.

Complete transcription conventions are provided in the appendix.All the names are pseudonyms.

2

3


T: Oh, this is the sheet? OK. Lemme just read it over a little bit (readssilently).. .Have you read the section on comparing and contrasting?

Writing center discourse generally includes the first two phases suggestedby Agar: diagnosis4 and directive, though the importance of the final phase,report, is relatively minor.5 Although the data from this study do not reveal thetight sequences reported by Thonus (1999a, 1999b), the basic phases are identi-fiable in most sessions. In fact, the first phase, diagnosis, appears to dominatewriting center sessions, a characteristic that is even more apparent in sessionswith L2 writers. A diagnosis can be offered in a very general sense by writersthemselves, as in the example (1) above. This is essentially a statement of theproblem, of what has brought the writer to the center, rather than a diagnosis ofwhat is wrong. In research on revision processes, this phase is often referred toas identification or detection and precedes diagnosis (Bereiter & Scardamalia,1987; Gumming & So, 1996; van Gelderen, 1997). The latter task almost alwaysfalls to the tutor, as in the excerpts below, both from sessions with L2 writers. Itmay be a global diagnosis, as in (2), in which the tutor points out where thewriter is moving in the wrong direction, or a local diagnosis, as in (3), in whichthe tutor focuses on an error in inflectional morphology.

(2) T (Eugenia); W (Evelyn, L1 = Chinese)

T: OK.but again, if you write that, that's going to be the same thing as thispaper.

W: What do you mean?T: That's not really what the teacher's looking for. He wants you to take

Willams and O'Rourke and tie them together and say why they are eitherthe same or why are they different. Or what is the correlation betweenthem?..Do you see what I'm saying?

W: mmhm.

(3) T (Eugenia); W (Sammy, L1 = Chinese)

W: More than three hundreds teachers from all over the country come toChicago, Illinois to participate in the show.

T: More than three hundreds teachersW: More than three hundreds teachers from all over the country come to

Chicago, Illinois to participate in the show.T: Ok, now let's look at..these two words...OK, Can 'hundreds' be a plural?W: mm.T: Can the word 'hundred' be a plural?W: ..Maybe, (laughs) ..No?T: No, it can't.

4 The use of the term diagnosis is perhaps unfortunate, given the widespread rejection of the medicalmetaphor in the writing center literature (e.g., North, 1984). However, since Agar's description ofinstitutional discourse has proven useful in other settings, I will pursue it here.5 Tutors submit a report after each session, and writers are also asked to provide a brief writtenevaluation of their experience. This phase is not part of the interaction, however.

42 Williams

In some cases, the diagnosis comes relatively early in the session, perhapsprompted by teacher comments on the draft. The structure of some sessions,especially ones with L2 writers that include some attention to sentence-levelissues, is a series of diagnosis-directive cycles. There are other instances inwhich the entire session is devoted to establishing a diagnosis. It is interesting tonote that in the current data, this happens only with L2 writers. In these cases,although the tutor may be able to detect surface or even more general problemsin the student's writing, it is not immediately clear what is at the root of theproblem. For the most part, the writers have either not understood theassignment or the readings they are required to incorporate into the assignment,as in this exchange (4) between a tutor and a L2 writer.

(4) T (Eugenia); W (Evelyn)

T: OK, so.in this paper, did you write about the similarities or the differences?W: I wrote um. Actually, I wrote, after I wrote this essay, then I reread it, then

I found out that.that.it.it doesn't meet the assignment (laughs)T: OK, so what did you write about in your essay?W: I wrote, I wrote Williams, Williams think there shouldn't be any racism and

..um.and then I gave some example why she thinks so.T: And then you also did PJ and the other side of it, right?W: but PJ, I did,...we should keep the bias in our language. And it doesn't.At

doesn't um.meet the assignment (laughs).

In this particular session, the diagnosis phase was quite lengthy. It took morethan 30 minutes to discover that that writer had completely misinterpreted one ofthe essays she was to analyze. She understood the author's statement that"neutrality hides racism" to be a recommendation of neutral language, ratherthan an indictment of bland politically correct language. Part of her misunder-standing was the result of her lack of familiarity with cultural institutions, suchas the purpose and process of writing letters to the editor, the widespread fear oflitigation in this country, and the function of a university law review. It is almostthree-quarters of the way through the session that the eureka moment occurred.The tutor had explained of the process of sending a letter to the editor of anewspaper and an editor's prerogative to edit the content of any letter submittedif it is deemed inflammatory. In fact, the author's (Williams') letter describing anincident of racial discrimination had been drastically cut for this very reason, butthe writer had not understood this. At the beginning of this excerpt, she is stillclinging to her original interpretation of the essay, that the author isrecommending the use of neutral language. By the end, after considerableexplanation by the tutor, she comes to realize that the two authors (Williams andO'Rourke), whose work she is comparing, have similar rather than contrastingviews.


(5) T (Eugenia); W (Evelyn)

W: But um, didn't we discuss earlier that Williams should keep neutral -keepum don't keep the bias in language? If the editor did that, then she will behappy about it?

T: OK, is this 'try to be neutral' the idea behind Williams or the idea behindher story?

W: Behind her story.T: OK, so it's not necessarily the way she feels, right? It's the way you see the

story...Do you see how these two aspects are different? Williams..! haven'tread this article, again.and I am just telling you, y'know, what I think fromthe information I have gotten from you. This is what I am understanding:that Williams um..wrote a very, like active, letter to the editor, expressingher feelings about how she was treated at the store, right? And it had all the..what the editor would call .like negative aspects.

W: mmhmT: In that it had racism and sexism and all the other y'know, her feelings in it.

Right? So what the editor tried to do is the editor sent her back the editedletter that was completely neutral, right? So the idea behind this letter, theeditor or whoever wanted to publish it, doesn't make it seem like it's areally big deal..by um..by making it simple, right? So the idea behind theletter is trying to be neutral hides racism. That's the idea behind the letter.

W: ..oh...OKT: Do you see how those two things can be different?W: uh huh (smiling) If this..if this idea is behind the story.behind um..um what

we came up with—that idea-then that would make sense.T: OK..OK, so now, how do you think this is related to O'Rourke?W: um hmT: OK, so O'Rourke says we've gone too far in trying to avoid bias in

language, right? OK, so do you think he is choosing the same side asWilliams? Which is a different side from the editor?

W: mmmmT: How do you think these two are similar or different?W: He said um..um. if he was the editor, he will publish her her essay.T: Alright! Great!

In fact, this session consists almost entirely of one long diagnosis sequence, withthe directive portion at the very end only implicit in the tutor's reinterpretation ofthe text. The L2 writer must infer that she should now revise based on this newinterpretation. Although this nondirective style was not typical in sessions withL2 writers, it was not uncommon. Indeed, writing center tutors are taught thatmuch of their job entails guiding writers to discover meaning for themselves, toidentify problems in their writing, and to figure out ways to improve it (e.g.,Brooks, 1991). Thus, the diagnosis phase is often a deliberately collaborativeand lengthy process, compared to say doctor-patient interaction, in which thedoctor does not generally invite the patient to participate in the diagnosis. Tightsequences of organized question-answer adjacency pairs, though represented inthe present data, are not the norm in writing center sessions. In excerpt (6), the

44 Williams

tutor tries to get the L2 writer to see that her thesis and introduction do not leadthe reader to the body paragraphs she has written. In this case, the teacher, in hercomments, had offered an initial diagnosis that the writer's paragraphs lackedlogical development, yet, again, it took almost the entire session to translate thisinto ideas that the writer could understand.

(6) T (Oscar); W (Min; L1 = Khmer)

W: This paragraph I just write.it's about .. .mm..telling about how he expl.howhe found that.his father told him the false stories.

T: mmhm. Right.now..um.so what did you wanna. Why did you include thatin your essay about finding out that his father told false stories?

W: ..m. to see.his father denies the past.T: OK.W: yeah.T: so.um. what's..why'd you include that in your introduction? How does that

relate to your introduction?..You just said..you just said his father wantedto talk about.I mean, the author wanted to talk about how his father—

W: —lost his identityT: mmm. Lost his identity .um..OK, so, which one is your thesis here in the

introduction?W: uh. My thesis is.about how his father lost.his father lost his identity..the

effect on his son.

Thus, one typical feature of writing center discourse is extended diagnosissequences, a characteristic accentuated in sessions with L2 writers.

THE STATUS OF INTERACTANTS

Institutional discourse structure also involves identification of participant roleand status, often designating on participant as dominant. However, as Thonus(1999a, 2001) and Ritter (2002) point out, once again, writing center sessionspresent a possible exception to, or at least a hybrid of this view. Although thetutors are consulted as relative experts, they are also presented as peers, osten-sibly as status equals, at least when interacting with NE writers. Indeed in anycontext, the marking of relative status is complex, and interactionally achieved.Yet, there are some structures and signals that are generally associated withexpert status that can begin to give a picture of the tutor's more dominant role inwriting center interaction (see Tarone, chapter 6, this volume). These are shownin Table 2.1.


Table 2.1. Markers and Conversational Moves Generally Associated with Expertise/Dominance

Dominance markers in interactionLonger turnsInterruptionsTopic initiation

Dominant speech actsSuggestions/requestsRejections/challenges

Lexical markers of dominanceLack of mitigation/modulationAggravators

Interactional Features of Writing Center Sessions That SuggestTutor Dominance

Turn Length. Tutor talk shows many signs of interactional dominance, afinding that corroborates those of earlier studies (Gumming & So, 1996; Ritter,2002; Thonus 1999a, 2001; Young, 1992). Among these are significantly longerturn length, as seen in Table 2.2. In general, tutors have longer turns thanwriters. There is a single exception in the data, in a session with a NE writer. Insessions with the L2 writers, the trend toward longer tutor turns is even stronger,suggesting greater dominance by tutors in their sessions with L2 writers.

Table 2.2. Turn Length in Number of Words

Writer-name

Esther-NE writerEsther-NE writerAnnie-NE writerAnnie-NE writer

Evelyn-L2 writerEvelyn-L2 writerMin-L2 writerMin-L2 writerSammy-L2 writerSammy-L2 writer

Tutor-name

JoeOscarRobertJoe

EugeniaEugeniaEugeniaOscarJoe ,Eugenia

Writer meanturn length

15.513.413.816.2

7.112.76.68.8

12.111.3

Tutor meanturn length

27.010.119.622.1

26.221.9

2716.829.622.2

Writer-tutordifference

6.5-3.35.85.9

19.19.2

20.48.0

17.5 10.9

46 Williams

Table 2.3. Interruptions by Tutors

Writer-name



Tutor-name

JoeOscarJoeRobert

EugeniaEugeniaEugeniaOscarJoeEugenia

Mean interruptionsper turn

.081

.071

.085

.182

.092

.088

.106

.077

.088

.119

Mean interruptions -tutor completes turn

.098

.100

.098

.063

.126

.136

.128

.131

.109

.111

Interruptions and Overlaps. Interruptions have also often been suggestedas a marker of speaker dominance (West & Zimmerman, 1985; Zuengler &Bent, 1991, though see Goldberg, 1990, for a summary of findings to thecontrary). The relative rate of interruptions in the writing center data showstutors wielding authority (Table 2.3). All of the figures in the table areinterruptions of the writer by the tutor. Interruption in the other direction, by thewriter, occurs very rarely, often only once or twice during a session andsometimes not at all.

It has been frequently noted that not all interruptions are necessarilydominating (Goldberg, 1990; Makri-Tsilipakou, 1994; Murata, 1994; Zuengler& Bent, 1991). Tannen (1989) and Murray (1985) have also disputed the equat-ing of interruption with speaker dominance, on both theoretical and methodo-logical grounds, yet there are clear cases in the data in which the tutor intrudesinto the writer's turn in mid-sentence. Column three of Table 2.3 contains inter-ruptions in which the learner cuts off the speech of the other interlocutor, eitherby changing the topic (7a), or dismissing or challenging the contribution (7b),moves that may be considered dominating. These have also been referred to asintrusive (Murata, 1994) or disaffiliative (Makri-Tsilipakou, 1994) interruptions.Among these, there is little difference between tutor interaction with NE writersand L2 writers. Differences seem to be more a matter of individual style or theproduct of the specific interaction between the two interlocutors.

(7a) T (Oscar); W (Min)

W: um..this one I already had a grade on it. But uh..I don't satisfy-I am notsatisfy with the grade I—

T: —What about the writing?


(7b) T (Oscar); W (Min)

W: This is uh..one pages. If you take out, it's uh--T: —No, no, I'm not saying take itout..but uh.. no, but um..how would this be a reflection of his father's changingbeliefs? Do you see what I'm asking now?

The last column in Table 2.3 includes interruptions that might be consideredmore collaborative or supportive.6 Tannen (1984, 1989) suggested that certaintypes of simultaneous speech demonstrate active participation and solidarityrather than dominance. It is possible that there is some effort on the part of thetutors to reduce social distance by engaging in overlapping speech; however, thepostsession interviews suggest a somewhat different interpretation. These ex-changes consist of instances in which the tutor finishes the writer's thought,often after at least a second of wait time. In many cases, the writer also chimesin with the end of the utterance, producing an overlap. These might be betterconsidered rescues than interruptions. In retrospective accounts, tutors reportedthat they were trying to guess what the writers intended to express and offeredthe utterance ending as a way of minimizing their struggle and embarrassment,especially in sessions with L2 writers (e.g., ex. 8). Writers, when questioned onthis point in the videotape, usually claimed that the interruption was helpful.

(8) T (Oscar); W (Min)

W: and then I use.the next paragraph is about the example about when he wentto the movies and he saw white people...

T: looking glamorous//W: //glamorous and perfect.

It is interesting to note that there are far more of these rescues of L2 writers thanof NE writers. Thus, it may be that these moves are at once dominating, in thatthey demonstrate the expertise of the tutors, and supportive, in that they attemptto follow the conversational aims of the writer. This is further evidence of thehybrid status of the tutor in writing center interaction.

Directives and Suggestions. Although the diagnosis phase of institutionaldiscourse is usually the lengthiest in writing center sessions, particularly in thosewith L2 writers, because most writers come to the writing center looking fordirection and advice, it is no surprise that almost all sessions include a directivephase as well. The directive phase of institutional interaction is one in whichexperts can demonstrate their authority and, as a result, their higher status. Theexperts make suggestions and give advice. The client or novice often collabo-rates to facilitate this dominance. Bardovi-Harlig and Hartford (1990) describe

6 Other terms for this type of interruption include cooperative (Murata, 1994) or affiliative (Makri-Tsilipakou, 1994).

48 Williams

situations in which contributions between status unequals must be congruentwith the roles that they recapitulate during interaction. In Bardovi-Harlig andHartford's work, faculty members advise students on course selections. It is thejob of the faculty to give advice. Students may make suggestions, or even rejectthose made by faculty, but they must do so in a way that appears congruent. Inconferences with writing teachers, this pattern of dominance has also beendemonstrated (Newkirk, 1995; Pathey-Chavez & Ferris, 1997; Sperling, 1991;Ulichny & Watson-Gegeo, 1989).

In the writing center, however, this is a more delicate situation as the statusof tutors is not always clear. According to accepted writing center practice,tutors are not supposed to appropriate student writing, or directly tell writers tomake changes in their writing (see e.g., Blau & Hall, 2002; Brooks, 1991;Capossela, 1998; Harris, 1982, 1995; Thonus, 1999b, 2001). In practice though,tutors often do give direct advice. And, as has been found repeatedly, this iswhat some writers, especially L2 writers, expect and want them to do (Blau &Hall, 2002; Clark, 2001; Harris, 1997; Thonus, 1999c; Young, 1992). All of theL2 writers in postsession interviews explicitly stated that their purpose incoming to the writing center was to have tutors make suggestions about how toimprove a specific piece of writing. Perhaps for cultural reasons, or simply outof politeness, few wished to voice their concern that they had not received asmuch advice as they had hoped. All but one remained noncommittal on theissue, saying that their tutor was "nice" or "helpful." The one writer withstrongly voiced complaints about the session was a NE writer who received theleast and least explicit advice on how to improve her paper. Instead, the tutorengaged the writer in a general discussion of the ideas that were developed inthe essay. The student was visibly disgruntled during the session, and in thepostsession interview declared that the tutor had been "no help," that he had"just rambled on and on."

There are many ways to make suggestions. Thonus (1999a, 1999c, 2002)addresses this issue first in terms of form: declarative, imperative, interrogativemodal (1st and 2nd person) and then goes on to examine how these all might bemitigated, based on work by Blum-Kulka, House, and Kasper (1989). However,this approach does not always capture the way in which mitigators or aggrava-tors can accumulate to mark status relations. In particular, modality can be usedeither to highlight the dominant status of the tutor or to mitigate it (e.g., youshould vs. you could) (He, 1993). Modal operators (should, might) and modaladjuncts (maybe, I think, conditionals, such as If you want) generally fulfill twofunctions: to mark logical possibility or to signal features of the social situationand interaction. In terms of their capacity to mitigate the potentially face-threatening act of a directive, only some modals are suitable. The followinggeneralization can be made: Modals of logical possibility that fall belowcertainty—might, may, probably, I think, maybe—have a mitigating or softeningeffect on directives. Modals of social interaction—should, have to, need to(though negated modals present a special case)—and modals of certainty—will,gonna—aggravate or heighten directives.


Modality is not the only way to adjust the force of directives. An inter-rogative tends to lessen the force of an utterance (Should we end it here!}whereas an imperative increases it (End it here.). The use of we tends to mitigatedirectives (Rounds, 1987; Thonus, 2002). The force of the directive can beupgraded with the addition of qualifiers such as really, more, and softened withchecks for agreement or comprehension, such as y'know?, right! (Rudolph,1994). Often these modulators are not used singly but in combination for acumulative effect. The following examples of tutor suggestions illustrate someof the possibilities, with mitigators and aggravators marked with subscripts:

(9a) T (Oscar)

T: Do you see how.places where you couldi maybe2 clear that up for me?3(Oscar)

(9b) T (Eugenia)

T: Okay, because 'studies' is really1 an awkward word to use here. You should2

use..m..

In the first example, the cumulative effect of the three signals is one ofmitigation, in the second, escalation. Of course, it is also possible that a directivecould contain modulation in both directions, and this does indeed occur. In thefollowing examples, the modulations with numeric subscripts mitigate the forceof the directive whereas those with the alphabetic subscripts augment them.

(10a)T (Eugenia)

The introduction is supposed1 to tell us everything wea are going to read about,rightf

(10b)T (Joe)

Once wea include your thesis at the end of. once wea include your thesis at theend of your introductory paragraph, then wea 're gonnat start talking about howwea came to your thesis.

Thus, consistent with their ambiguous status, tutors often do a delicate dance ofexerting authority and reducing status difference though their linguistic choices.Table 2.4 shows tutor suggestions in terms of the degree to which their force iseither mitigated or heightened through the use of the devices described earlier.A score of 1+ means that the suggestion contains one aggravator; 2+, two aggra-vators. A score of 1- means one mitigator and so forth. "Mixed" means that thesuggestion contained both types of modulators.

Table 2.4. Modulation of Tutor Suggestions

Sug-Tutor gestions

N

NE writers

JoeOscar

RobertJoe

Total NE writers

L2 writers

EugeniaEugeniaEugenia

OscarJoe

Eugenia

Total L2 writers

49393847

173

273239232834

183

1+N

5526

75

8797

1014

55

1+%

10.212.85.3

12.8

10.4

29.621.923.130.435.741.2

30.1

2+N

1000

1

323023

13

2+%

2.0000

0.6

11.16.37.7

07.18.8

7.1

1+N

128

1711

48

1012178

1310

70

1-%

24.520.544.723.4

27.7

37.037.543.634.846.429.4

38.3

2-N

19181419

70

454513

22

2-%

38.846.236.840.4

40.5

14.815.610.321.73.68.8

12

3-N

5436

75

010000

1

3-%

10.210.37.9

12.8

10.4

03.1

0000

.05

MixN

7425

75

245312

77

Mix%

14.310.35.3

10.6

10.4

7.412.512.813.03.65.9

9.3

OtherN

0000

0

0010

0120

4

Other%

0000

0

00

2.60

3.65.9

2.2


A writer may also request advice from the tutor. Although this does put thewriter in the dominant position of controlling the flow of discourse, this positionis immediately relinquished when the tutor begins to offer advice. Furthermore,writers usually only make these requests when authorized to do so by the tutor.

(11)T (Joe); W (Esther)

T: How can I help you today?W: Well, I have to write.like.a compare and contrast paper on religion. How

two authors have wrote.and I have to write a paper about.what I thinkum.whether or not we need religion....um.but I'm kind of stuck. I don'tknow really how to write a compare and contrast paper. / need some helpwith that.

Only in rare instances does a writer make an unlicensed request. In the followingexcerpt, the writer is quickly cut off when she does so and is told that it is notthe job of tutors to give direct advice. In fact, however, this tutor had done sorepeatedly in this and other sessions.

(12) T (Eugenia); W (Min)

W: oh..mmni.cause I just can't say 'disadvantage' because 'disavantage' is justlike < > Just like one point—

T: —OK. Um, let's see if we can look up a similar word for 'disadvantage'(looks up in dictionary) [softly] It's not in here (puts book back).

W: Can you just give me some like—T: —Well, I can't really give [loud] you an idea. Um. How'bout you think of

a title for it?W: Title, yeah. Can you give me a title this column?T: I can't really give you a title (laughs).R: OK.

Or more gently:

(13) T (Eugenia); W (Min)

W: Today I would like to work on my grammar.T: Okay.W: um...Just my grammar..um..and..and please tell me if there's any..um..if

there's any area I need to explain so I can have a clearer idea—T: —Okay. Well, let's work on that first and then we can put off the grammar

until like..the last ten or fifteen minutes.W: uh huh, okay.

Here, ironically, the tutor uses the nondirective philosophy of the center tomaintain her authoritative stance.

52 Williams

Topic Shifts/Leading Moves. The dominant interlocutor controls the flowof the discourse, pursuing or shifting topics. Topic shifts are difficult to quan-tify, or sometimes even identify in writing center sessions because, in somecases, the tutor and writer seem to pursue the same topic the whole time.Therefore, the analysis will focus specifically on moves that push the interactionforward, what Gass and Varonis (1986) call leading moves. Although it ispossible for an interlocutor to push the conversation forward in a more implicitfashion, for instance, with body movement or gaze direction, only explicitmoves in which the movement forward is marked linguistically, are countedhere (14a,b).

(14a)T (Joe); W (Sammy)

T: I think that pretty much is going to improve your um..first page a lot more.Is there anything else you want to work on the first page on?

(14b)T (Eugenia); W (Eveyln)

T: OK..um. let's put off the question for a second. How bout O'Rourke? Whatwas his story about?

Table 2.5 shows that tutors use far more leading moves than writers. Again, thetutor dominance is evident in their control over the flow of the session, thoughthere is no clear difference between their interaction with L2 and NE writers.

Rejections and Challenges. Rejections and challenges are risky and face-threatening acts (Gass & Houck, 1999). Learning to challenge suggestions madeby others, especially those of potentially higher status takes more thanknowledge of the language. It has been amply demonstrated that grammaticalcompetence is not a guarantee of pragmatic competence (Bardovi-Harlig, 2001;Bardovi-Harlig & Dorrnyei, 1998) and that developing this competence takes along time (Bardovi-Harlig & Hartford, 1993b; Salsbury & Bardovi-Harlig,2000). Because of the status difference between the participants, it is notsurprising that challenges are relatively rare in the data. When the tutor offersadvice, generally it is either accepted or briefly acknowledged (though in theend, often ignored). Only rarely do writers explicitly reject a tutor's suggestionor request, and when they do, they use some form of mitigation, such asdowngraders or an excuse or explanation (see Bardovi-Harlig & Hartford, 1991;Gass & Houck, 1999). Tutor response varies from acceptance of the challenge togentle rebuff to mild anger. In (13) the tutor offers the highly mitigated(mitigation in italics) suggestion that one paragraph in the writer's essay may notfit. The writer rejects the suggestion that it be cut with explanation that theresulting essay will be then too short to fulfill the assignment. The tutor backsoff and revises the suggestion.


Table 2.5. Leading moves

Tutor-name

JoeOscarJoeRobert


Leadingmoves

4635

464343

Writer-name



Leadingmoves

1101

000100

(15)T(Oscar);W(Min)

T: Because that's the only one. Either that one sticks out or this one sticks out,right? But it's about..if you're talking about um.. changing.changingbeliefs, you said?

W: mmhmT: Awright. I guess the one paragraph that sticks out if you're talking about

changing beliefs is this first one, right? Or no? Or with the father at least.W: This is uh..one pages. If you take out, it's uh—T: —No, no, I'm not saying take it out..but uh..no, but um..how would this be

a reflection of his father's changing beliefs? Do you see what I'm askingnow?

W: mmhm mhhmm... yah...

Excerpt (16) contains a rejection of a request rather than of a suggestion.Earlier in this session, the tutor had asked the writer to read aloud, but shereturned with a request that he read it instead. Here, he presses her, not accept-ing her excuse of poor pronunciation. He does, however, provide a justificationfor his rejection of her initial challenge, that he needs her to read it aloud inorder to evaluate it.

(16)T(Oscar);W(Min)

T: OK..um..Actually, I need to hear it...It helps me to think. Can you—W: BecauseT: Or does it take—W: I get so..It's get a lot from .1 get a lot of evidence from the book and some of

the evidence, I don't know how to pronounce.T: Oh, that's fine.W: Can you?T: When you get to that part, you can just-J'll read it to myself, but I want to

hear you read your writing.. .OK?

54 Williams

In the final example of challenges, (15), the writer had repeatedly met thetutor's suggestions for changes in grammatical form and word choice—the focusof the session at his request—with the response "maybe," a modal expressionthat Salsbury and Bardovi Harlig (2000) found was favored by low-proficiencyL2 learners in challenges and refusals. In this excerpt, toward the end of thesession, the tutor loses patience. In fact, this is the only writer in the corpus whooffers barely mitigated rejections, perhaps serving as a challenge to the tutor'sauthority. This may explain the tutor's somewhat explosive response.

(17) T (Eugenia); W (Sammy)

T: See, you did it right here. It is more than just a noun, right?W: yeahT: So you should have done the same thing here, right?W: Maybe.T: [loud] What do you mean, maybe? Maybe? If you take out the "more"

here, "It is just a noun"..right? It's the same sentence as you have righthere.

W: mhm.T: OK, go on.

In the postsession discussion, this writer was described by his tutor asdifficult and uncooperative. She expressed frustration that after she hadacquiesced to his request to go over grammatical problems in his paper, againsther better judgment, he seemed to show little willingness to accept her advice."Why did he bother to ask if he didn't want my advice?" she asked. It was herview that he had agreed to come to a session simply in order to collect thepayment offered for participation in the project. Indeed, of the writers whoparticipated in the study, he is the only one who did not return to the writingcenter after the project ended.

Negotiating Interactional Status

Floor Management. So far, our primary focus has been how tutors showinteractional dominance. However, it has been pointed out that because statuscan be interactionally achieved, the nondominant party in interaction oftencollaborates to allow the dominant party to lead and to hold the floor (Dyehouse,1999; Rudolph, 1994). One way in which this can be achieved is through back-channeling. This practice shows that the listener is interested and attending butis not attempting to claim the floor.7 Like the leading moves, these were codedconservatively, with only audible signals counted. Most instances are singleword utterances (yeah, OK), or simply verbalizations such as uh huh or mmm.These are considered backchannels when they follow the other speaker's turn

7 In fact, however, backchanneling, especially minimal signals by L2 writers, may not be signal ofcomprehension or agreement. Williams (in press) demonstrates that tutor suggestions that meet withminimal signals of L2 writer attendance are rarely taken up in subsequent revisions.


with little or no pause, and there is no subsequent speech from that speaker.Often the turns are simultaneous, that is, the first speaker continues with the turnwhile the back channel is offered.

Here again we see that the writers regularly yield to tutors (Table 2.6).However, there is no clear trend differentiating NE writers from L2 writers;instead, there is more individual variation across pairs. Among the tutors, forexample, Joe backchannels very little whereas Oscar does so extensively, onceeven more often than the writer with whom he was working. Both of these styleshold across writer partners. Sammy, the writer described above, who displays avariety of uncooperative behaviors, shows the lowest level of backchanneling ofall of the writers. It is important to underscore that though the writersconsistently occupy the nondominant position in these interactions, they areactive participants in constructing this role for the tutors.

We have seen the tutors generally dominate the interaction in the writingcenter sessions, that they more freely offer face-threatening speech acts, such assuggestions and advice, that most backchannel less than the writers, and thatthey hold the floor for longer. All of this suggests they perceive themselves andthat the writers perceive them as experts-authorities rather than peers, much ashas been described in work on institutional interaction between experts andclients. In the case of writing center tutors, however, roles can be ambiguous andgoals become fuzzy. Thus, in addition to markers of interactional dominance, wealso see moves that may be intended to decrease status difference and increasesolidarity with the writer. In addition to displaying their expertise and authority,tutors may try to establish comembership to show that they are status equalswith their interactants. We have already seen some of this in the backchannelingbehavior of tutors and the mitigation of face-threatening speech acts.

Small Talk. Tutors may also try to reduce status differences and personal-ize the interaction with small talk (Erickson & Schultz, 1982; Thonus, 2002) andthe establishment of common experience and of a personal stake in the outcome

Table 2.6. Backchanneling

Tutor-name

JoeOscarJoeRobert


Backchannelcues

8696

37

9141721

317

Writer-name



Backchannelcues

46593452

39596147

824

56 Williams

of the process. This can establish comity (Aston, 1993) or more specifically,what Bardovi-Harlig and Hartford (1993) describe as social comembership. Thiscontrasts with role comembership, which tutors and writers cannot claim sincethe institution has assigned them different roles. Again, there is considerablevariation across tutors, with some more comfortable in an authoritative role andothers gravitating toward a peer relationship. In the first excerpt (18a), the tutorempathizes with the writer's situation, calling on their common student status. Inthe second (18b), the tutor alludes to their shared status as Asian-Americans.

(18a)T (Oscar); W (Esther)

W: I just finished my paper so I'm like—T: yeah, you 're kickin' back.W: yeah.. .1 should start.T: And you don't—so this is kind of hard for you to talk about it now? Cause

you wanna relax?W: yeah.

(18b)T (Oscar); W (Esther)

W: And they got a little bit more freedom than me and sometimes I waslike.you know.oh.just because I'm Korean, I'm like this.

T: mmhmW: because I'm Korean, I can't go out as much..other people—T: —It's also because you're a girl. If you were a Korean boy, they probably

would let you.W: yeah.also like when I go like when I come come in a little bit past my

curfew. My parents will be like, if you were a—like again-T: —oh, what's your curfew?W: Curfew? Curfew is like 12:00.T: Oh, OK, that's not different. Cause my friends have like 10:00.

In the next example (19), the tutor's strategy for reducing status differencesis a little different. In this exchange, he develops a small talk sequence todemonstrate a personal interest in her cultural heritage, specifically, Koreanfood. He pursues this topic, providing information about his own personalexperience and preferences, even when the writer displays little inclination topursue the topic.

(19) T (Oscar); W (Esther)

T: But what about when like when they came over..and they didn't knowKorean? Did they ever come over?

W: um..yeah..they came over and...it was weircLlike I tried to show them likeKorean I tried to tell them about like.you know.my Korean. Sometimesthey were like they would be like.they wouldn't really understand.

T: What about the food? (laughs)W: yeah..They wouldn't understand.. .Yeah, the food is different, too..Um..


T: I like it.W: yeahT: [softly] It's great (laughs)W; oh yeah.Korean food?T: Only, it's only thing I know is-I can't pronounce it..forget it.W: OKT: the meatW: bulgogi?T: yeah! That one!

In some cases, this contradiction between peer and expert status can beconfusing. In this final excerpt (20), the tutor senses—perhaps signaled by thewriter's minimal response—that he has breached the normal constraints of thesession by providing too much personal information, and he apologizes.

(20) T (Oscar); W (Min)

T: Mkay urn..Ok well, are there any physical details about the house.thatyou.that you

W: mmT: or you can't think of that?W: it's uh.. .physical [softly]T: Anything that.well, if it doesn't stick out in your head, then maybe, y'know,

you can leave it alone.W: OKT: like... my house has red carpet (laughs)W: ohT: yeah, my parents like red, so we have red furniture, red carpet, a red car,

so I like, that's what I would write about.W: mm..OKT: or pets and plants-we have a lot.well not pets, we have a lot of plants.W [softly] ohT: Ok, I'm sorry. Let's keep going...um so we go on faith.

COMPARABILITY OF INTERACTION WITH NE AND L2 WRITERS

We have already noted that some characteristics of tutor-writer interaction aremore prominent in sessions with L2 writers: Both the diagnosis phase and tutorturn length tends to be longer in interaction with L2 writers. In addition, thenature of tutor interruptions of NE and L2 writers differs somewhat, with tutorsmore likely to make supportive interruptions that "rescue" L2 writers. One otherclear trend, shown in both quantitative and qualitative terms, is apparent in thedirectives and suggestions made by tutors. First, the density of advice is fargreater in NE writer sessions than in L2 writer sessions (Table 2.7). Whenquestioned about the smaller number of suggestions or advice in their sessionswith L2 writers, two of the tutors maintained that they wanted to make sure thatthe writer understood and concentrated on the main issues, and that they did notwant to confuse the writers by "giving them too many ideas or directions." It is

58 Williams

Table 2.7. Tutor Suggestions

Writer-name



Tutor-name

JoeOscarJoeRobert


Number oftutor turns

142138102141

118162181129156167

Number of tutorsuggestions

49394733

273239232834

Mean suggestionsper turn

.35

.28

.46

.23

.23

.20

.22

.18

.18

.20

Note: These are suggestions for improving the text, and do not include directives such as "Pleaseread me your essay."

important to note the differences between NE writer and L2 writer sessionscannot be attributed exclusively to what one might call "grammar help." ManyL2 writers go to the writing center for help with second language issues, whichthey usually refer to as "grammar." Of course, NE writers also often request helpwith sentence-level correctness. However, since the philosophy of the writingcenter is to work on the writing process, not simply the surface accuracy of theproduct, tutors are often inclined to deflect or delay any discussion of grammar.In fact, only one session in this corpus addressed grammar issues as a primaryfocus, at the insistence of the writer. There are references to grammatical issuesin most of the sessions with NE writers and L2 writers alike, but they do notconstitute a major part of the sessions.

Second, there are qualitative differences in tutor interaction with NE writersand with L2 writers. In contrast to Thonus' findings, tutors in this study aregenerally more direct in their suggestions to L2 writers than to NE writers. Thedifferences across suggestion types for NE writers and L2 writers are consider-able: Just 12.6% of the suggestions made to NE writers are heightened, incontrast to 37.2% of those made to L2 writers; 75.6% of the suggestions made toNE writers are mitigated, compared to 55.3% for the L2 writers. In addition, L2writers seem to get simpler suggestions, with fewer modulations of any kind.Figure 2.1 shows the percentage of suggestions made with one, two and threemodulations, combining the mitigators and aggravators. It shows that sugges-tions to L2 writers tend to be simpler than those aimed at NE writers.

Bardovi-Harlig and Hartford (1993a) suggest that more elaborate encodingscome in interaction between speakers of closer status, in which roles and author-ity have to be negotiated, whereas status unequals are likely to be more indirect,consistent with the Bulge Theory, originally suggested by Wolfson (1988). Ifone argues that native speaker status brings these writers closer into comember-ship with their tutors, then these findings are consistent with those of Bardovi-


Figure 2.1. Number of Modulations in Tutor Suggestions

Harlig and Hartford. With L2 writers, tutors take on the authority of the statusmore easily, displaying it more frequently through unmitigated or more forcefulsuggestions. Another possibility is pointed out by Thonus (1999b), that tutorsuse these simpler modulation strategies in an effort to increase the compre-hensibility of their suggestions to L2 writers. It may be that the multiple use ofupgraders, though perhaps less polite, serves comprehensibility. Tutors are farmore likely to use at least one upgrader with the L2 writers than with NEwriters, and even more likely to use two. This may make tutor directives clearer.This interpretation is consistent with post-session interviews with tutors. Whenthey were questioned about the form of their suggestions, at first, most seemedunaware that they were behaving any differently with L2 writers than with NEwriters. On reflection, two tutors thought that they might have used simpler andmore direct and unmitigated suggestions because they thought that the writerwould be more likely to understand the advice expressed in this way andconsider it seriously.

CONCLUSIONS AND IMPLICATIONS FOR SLA

The structure of the writing center sessions clearly reflects patterns and featuresreported for institutional discourse. It has a predictable structure; it has set rolesfor participants and a narrow range of topics. Within this generalization, there issome variation. Although almost all sessions contain goal-setting, diagnosis anddirective phases, sessions with L2 writers tend to favor a very lengthy diagnosis

60 Williams

stage, perhaps as the interlocutors struggle to come to a mutual understanding oftheir task. The range of permissible topics is limited. For example, although asmall amount of initial and closing small talk is tolerated and even expected, if aparticipant ventures too far or too long outside of the established range of topics,the other participant usually reigns him in with signals of discomfort or lack ofinterest in response.

Roles are negotiated during the sessions, but again, only within a predict-able range. Tutors' interactive style blends characteristics of dominance andsolidarity, underscoring their dual roles as experts and peers relative to thewriters they are assisting. Tutors must consistently maintain this balancebetween authority and equality. One possible way of describing their role isoffered by Vygotsian approaches to education, that of the more capable peer(Moll, 1989). A growing body of research into peers assisting one other in theconstruction and internalization of new knowledge has demonstrated the benefitof collaboration with true peers (Donato, 1994; Storch, 2002; Swain, Brooks, &Tocalli-Beller, 2002; Swain & Lapkin, 1998; Villamil & De Guerrero, 1998; seeLantolf, 2000, for a review). Collaboration with writing center tutors issomewhat different, and potentially even more beneficial since they lay claim tosome expertise and authority as well as commonality of experience with thewriters, hence the term, more capable peer. Writing center interaction is consis-tent with what Storch (2002) calls an expert-novice pattern, in which oneinterlocutor (the tutor) generally controls the flow of discourse, demonstrating alack of equality. Yet, there is moderate mutuality, that is, the expert activelyencourages the participation of the novice.

Though characteristics of both dominance and solidarity are present in thewriting center sessions, the data here suggest that both participants willinglytake on nonreciprocal roles in their interaction. The balance is toward tutordominance and authority, as shown through turn length, floor management, andtheir use of potentially face-threatening speech acts. This trend is morepronounced in sessions with L2 writers on many, but not all measures. Althoughthis tendency toward dominance may be linked to their native speakerknowledge, it is not generally manifested in a display of linguistic knowledge.That is, the content of the sessions does not usually emphasize their superiornative speaker competence, so this is not a compelling explanation for thedifference between L2 writer and NE writer sessions. One possible explanationagain, may be the effort on the part of the tutors to facilitate L2 writercomprehension. This may also be a factor in the simpler and more explicitdirective forms that the tutors use with these students. Whether or not thisinteractive style is beneficial for the L2 writers remains to be seen.

Finally, there are two areas in which these findings may shed light on thedevelopment of SL pragmatic competence. First, given the somewhat differentstances of the tutors with NE and SL writers, we may wish to question thevalidity of using NS speech events in developing target norms for SL learners.Do NE and SL writers participate in equivalent speech events, or are thereimportant differences that impact norms for interaction? Do the norms change as


a learner's proficiency increases? In most experimental studies of interlanguagepragmatics, there is an underlying assumption that the speech situation isconstant; SL performance is measured against that of NSs in the same speechsituations, an assumption that extends to assessment of pragmatic competence(e.g., Brown, 2001; Hudson, 2001). Yet, in writing center interaction at least,this assumption is only partially supported; indeed the input to L2 learners,suggested as a crucial factor in L2 pragmatic development (Bardovi-Harlig,2001), may differ from the input in NE sessions. Second, the findings under-score the importance of learning not just rules for taking on specific roles ininteraction, such as a graduate student in an advising session or client in a socialservice interview, but of learning how these roles can be actively negotiated ininteraction. This presents a considerably more complex goal for the acquisitionof SL pragmatic competence.

ACKNOWLEDGMENTS: This research was supported by a grant from the University of Illinois atChicago Office of Vice Chancellor for Research. My thanks to Leane Dostaler for assistance intranscription and coding, and to the editors for their valuable suggestions.

REFERENCES

Agar, M. (1985). Institutional discourse. Text, 5, 147-168.Aston, G. (1993). Notes on the interlanguage of comity. In G. Kasper & S. Blum-Kulka (Eds.),

Interlanguage pragmatics (pp. 224-250). Oxford, England: Oxford University Press.Aleksa, V., Bednarowicz, E., Smith, E., Brecke, C., & Huang, G. (2000). Between a better grade

and a meaningful conversation: Departures for writing interlocutors. University of Illinois atChicago Writing Center handbook.

Bardovi-Harlig, K. (2001). Evaluating the empirical evidence: Grounds for instruction in pragmatics.In G. Kasper & K. Rose (Eds.), Pragmatics in language teaching (pp. 13-32). Cambridge,England: Cambridge University Press.

Bardovi-Harlig, K., & Dornyei, Z. (1998). Do language learners recognize pragmatic violations?Pragmatic versus grammatical awareness in instructed L2 learning. TESOL Quarterly, 32, 233-259.

Bardovi-Harlig, K., & Hartford, B. (1993a). The language of comembership. Research on Languageand Social Interaction, 26, 227-257'.

Bardovi-Harlig, K., & Hartford, B. (1993b). Learning the rules of academic talk: A longitudinalstudy of pragmatic development. Studies in Second Language Acquisition, 15, 279-304.

Bardovi-Harlig, K., & Hartford, B. (1991). Saying "no" in English: Native and normative rejections.Pragmatics and Language Learning, 2, 41-57.

Bardovi-Harlig, K., & Hartford, B. (1990). Congruence in native and normative conversations:Status balance in the academic advising session. Language Learning, 40, 467-500.

Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ:Lawrence Erlbaum Associates.

Blau, S., & Hall, J. (2002). Guilt-free tutoring: Rethinking how we tutor non-native-English-speaking students. Writing Center Journal, 23, 23-44.

Blau, S., Hall, J., & Strauss, T. (1998). Exploring the tutor-client conversation: A linguistic analysis.Writing Center Journal, 19, 19-48.

Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-cultural pragmatics: Requests andapologies. Norwood, NJ: Ablex.

Bouquet, E. (1999). 'Our little secret:' A history of writing centers, pre-to post-open admissions.College Composition and Communication, 50, 462-483.

62 Williams

Brooks, J. (1991). Minimalist tutoring: Making students do all the work. Writing Lab Newsletter, 15,1-4.

Brown, J.D. (2001). Pragmatic tests: Different purposes, different tests. In G. Kasper & K. Rose(Eds.), Pragmatics in language teaching (pp. 301-325). Cambridge, England: CambridgeUniversity Press.

Bruffee, K. (1984). Peer tutoring and the conversation of mankind. In G. Olsen (Ed.), Writingcenters: Theory and administration (pp. 3-14). Evanston, IL: NCTE.

Cameron, R., & Williams, J. (1997). Sentence to ten cents: A case study of relevance andcommunicative success in normative-native speaker interactions in a medical setting. AppliedLinguistics, 18, 415-445.

Capossela, T.-L. (1998). The Harcourt Brace guide to peer tutoring. Fort Worth: Harcourt Brace.Clark, I. (2001). Perspectives on the directive/non-directive continuum in the writing center. Writing

Center Journal, 22, 33-57.Gumming, A., & So, S. (1996). Tutoring second language text revision: Does the approach to

instruction and the language of communication make a difference? Journal of SecondLanguage Writing, 5, 197-225.

Damon, W., & Phelps, E. (1989). Critical distinctions among three approaches to peer education.International Journal of Educational Research, 58, 9-19.

Donate, R. (1994). Collective scaffolding in second language acquisition. In J. Lantolf & G. Appel(Eds.), Vygotskian approaches to second language research (pp. 33-56). Norwood, NJ: Ablex.

Drew, P., & Heritage, J. (1992). Analyzing talk at work. In P. Drew & J. Heritage (Eds.), Talk atwork (pp. 3-65). Cambridge, England: Cambridge University Press.

Dyehouse, J. (1999). Peer tutors and institutional authority. In L. Podis & J. Podis (Eds.), Workingwith student writers (pp. 53-57). New York: Peter Lang.

Erickson, F., & Schultz, J. (1982). The counselor as gate-keeper. New York: Academic Press.Fiksdal, S. (1989). Framing uncomfortable moments in cross-cultural gatekeeping interviews. In S.

Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition:Discourse and pragmatics (pp. 190-207). Clevedon, UK: Multilingual Matters.

Fisher, S., & Todd, D. (Eds.). (1983,). The social organization of doctor—patient communication.Washington, D.C.: Center for Applied Linguistics.

Gadbow, K. (1992). Foreign students in the writing lab: Some ethical and practical considerations.Writing Lab Newsletter, 17, 1-5.

Gass, S., & Varonis, E. (1986). Sex differences in normative speaker-normative speaker interactions.In R. Day (Ed.), Talking to learn (pp. 327-351). Rowley, MA; Newbury House.

Gass, S., & Houck, N. (1999). Interlanguage refusals: A cross-cultural study of Japanese-English.Berlin: Mouton de Gruyter.

Gass, S., & Mackey, A. (2000). Stimulated recall methodology in second language research.Mahwah, NJ: Lawrence Erlbaum Associates.

Gillespie, P., Gillam, A., Brown, L.F., & Stay, B. (Eds.) (2002). Writing center research: Extendingthe conversation. Mahwah, NJ: Lawrence Erlbaum Associates.

Goldberg, J. (1990). Interrupting the discourse on interruptions. Journal of Pragmatics, 14, 883-903.Harris, M. (1982). Tutoring writing: A sourcebookfor writing labs. Glenview, IL: Scott Foresman.Harris, M. (1995). Talking in the middle: Why writers need writing tutors. College English, 57, 27-

42.Harris, M. (1997). Cultural conflicts in the writing center: Expectations and assumptions of ESL

students. In C. Severino, J. Guerra, & J. Butler (Eds.), Writing in multicultural settings (pp.220-233). New York: Modern Language Association.

Harris, M., & Silva. T. (1993). Tutoring ESL students: Issues and options. College Composition andCommunication, 44, 525-537.

He, A. (1993). Exploring modality in institutional interactions: Cases from academic counsellingencounters. Text, 13, 503-528.

He, A. (1994). Withholding academic advice: Institutional context and discourse practice. DiscourseProcesses, 75,297-316.

Hudson, T. (2001). Indicators for pragmatic tests: Some quantitative tools. In G. Kasper & K. Rose(Eds.), Pragmatics in language teaching (pp. 283-300). Cambridge, England: CambridgeUniversity Press.


Kasper, G., & Rose, K. (1999). Pragmatics and SLA. In W. Grabe (Ed.), Annual Review of AppliedLinguistics, 19, 81-104.

Kennedy, B. (1993). Non-native speakers as students in first-year composition classes with nativespeakers: Can writing tutors help? Writing Center Journal, 13, 27-38.

Kinkead, J., & Harris, J. (Eds.) (1993). Writing centers in context. Urbana, IL: NCTE.Lantolf, J. (2000). Second language learning as a mediated process. Language Teaching Abstracts,

33, 79-96.Lave, J., & Wenger. E. (1991). Situated learning: Legitimate peripheral participation. Cambridge,

England: Cambridge University Press.Makri-Tsilipakou, M. (1994). Interruption revisited: Affiliative vs. disaffiliative interruption.

Journal of Pragmatics, 21,401-426.Mehan, H. (1985). The structure of classroom discourse. In T. van Dijk (Ed.), Handbook of

discourse analysis, Vol. 3 (pp. 120-132). London: Academic Press.Moll, L. (1989). Teaching second language students: A Vygotskyan perspective. In D. Johnson & D.

Roen (Eds.), Richness in writing (pp. 55-69). New York: Longman.Moser, J. (1993). Crossed currents: ESL students and their peer tutors. Research and Teaching in

Developmental Education, 9, 37-43.Murata, K. (1994). Intrusive or co-operative? A cross-cultural study of interruption. Journal of

Pragmatics, 21, 385-400.Murphy, S., & Law, J. (Eds.). (1995). Landmark essays on writing centers. Mahwah, NJ: Lawrence

Erlbaum Associates.Murray, S. (1985). Towards a model of members' methods for recognizing interruptions. Language

in Society, 73,31-41.Newkirk, T. (1995). The writing conference as performance. Research in the Teaching of English,

29, 193-215.North, S. (1984). The idea of a writing center. College English, 46, 433-446.Pathey-Chavez, G.G., & Ferris, D. (1997). Writing conferences and the weaving of multi-voiced

texts in college composition. Research in the Teaching of English, 31, 51-90.Powers, J. (1993). Rethinking writing center conferencing strategies for the ESL writer. Writing

Center Journal, 13, 39-47.Ritter, J. (2002). Negotiating the center: An analysis of writing center tutorial interactions between

ESL learners and native-English speaking writing center tutors. Unpublished doctoraldissertation, Indiana University of Pennsylvania..

Rounds, P. (1987). Characterizing successful classroom discourse for NNS teaching assistanttraining. TESOL Quarterly, 21, 643-671.

Rudolph, D. (1994). Constructing an apprenticeship with discourse strategies: Professor-graduatestudent interactions. Language in Society, 23, 199-230.

Salsbury, T., & Bardovi-Harlig, K. (2000). Oppositional talk and the acquisition of modality in L2English. In B. Swierzbin, F. Morris, M. Anderson, C. Klee, & E. Tarone (Eds.), Social andcognitive factors in second language acquisition (pp. 57-76). Somerville, MA: CascadillaPress.

Sarangi, S., & Roberts, C. (Eds.). (1999). Talk, work and institutional order. Berlin: DeGruyter.Sperling, M. (1991). Dialogues of deliberation: Conversation in the teacher-student writing

conference. Written Communication, 8, 131-162.Storch, N. (2002). Patterns of interaction in ESL pair work. Language Learning, 52, 119-158.Swain, M., Brooks, L., & Tocalli-Beller, A. (2002). Peer-peer dialogue as a means of second

language learning. Annual Review of Applied Linguistics, 22, 171-185.Swain, M., & Lapkin, S. (1998). Interaction and second language learning: Two adolescent French

immersion students working together. Modern Language Journal, 83, 320-338.Tannen, D. (1984). Conversational style: Analyzing talk among friends. Norwood, NJ: Ablex.Tannen, D. (1989). Interpreting interruption in conversation. Chicago Linguistics Society Papers, 25,

266-287.Tannen, D., & Wallat, C. (1993). Interactive frames and knowledge schemas in interaction:

Examples from a medical examination/interview. In D. Tannen (Ed.), Framing in discourse(pp. 57-78). Oxford, England: Oxford University Press.

64 Williams

Tarone, E., & Kuehn, K. (2000). Negotiating the social services oral intake interview. TESOLQuarterly, 34, 99-126.

Thonus, T. (2002). Tutor and student assessments of academic writing tutorials: What is "success?"Assessing Writing, 8, 110-134.

Thonus, T. (2001). Triangulation in the writing center: Tutor, tutee, and instructor perception of thetutor's role. Writing Center Journal, 22, 59-81.

Thonus, T. (1999a). Dominance in academic writing tutorials: Gender, language proficiency and theoffering of suggestions. Discourse and Society, 10, 225-248.

Thonus, T. (1999b). How to communicate politely and be a tutor, too: NS-NNS interaction andwriting center practice. Text, 19, 253-279.

Thonus, T. (1999c, March). NS-NNS interaction in academic writing tutorials: Discourse analysisand its interpretations. Paper presented at the annual conference of the American Associationof Applied Linguistics, Stamford, CT.

Tyler, A. (1995). The coconstruction of cross-cultural miscommunication: Conflicts in perceptions,negotiation and enactment of participant roles and status. Studies in Second LanguageAcquisition, 17, 129-152.

Trimbur, J. (1987). 'Peer tutoring': A contradiction in terms? Writing Center Journal, 7, 21-28.Ulichny, P., & Watson-Gegeo, K. (1989). Interactions and authority: The dominant interpretive

framework in writing conferences. Discourse Processes, 12, 309-328.van Gelderen, A. (1997). Elementary students' skills in revising. Written Communication, 14, 360-

397.Villamil, O., & De Guerrero, M. (1998). Assessing the impact of peer revision on L2 writing.

Applied Linguistics, 19, 491-514.West, C., & Zimmerman, D. (1985). Gender, language and discourse. In T. van Dijk (Ed.),

Handbook of discourse analysis, Vol.4 (pp. 103-124). London: Academic Press.Williams, J. (in press). Tutoring and revision: Second language writers in the writing center. Journal

of Second Language Writing.Woken, M., & Swales, J. (1989). Expertise and authority in native-nonnative conversations: The

need for a variable account. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variationin second language acquisition, Vol.1 (pp. 211-222). Clevedon, UK: Multilingual Matters.

Wolfson, N. (1988). The bulge: A theory of speech behavior and social distance. In J. Fine (Ed),Social language discourse: A textbook of current research (pp. 21-38). Norwood, NJ: Ablex.

Young, Virginia. Politeness phenomena in the university writing conference. Unpublished doctoraldissertation, University of Illinois at Chicago. 1992.

Zuengler, J., & Bent, B. (1991). Relative knowledge of content domain: An influence on native-nonnative conversations. Applied Linguistics, 72,397-416.


APPENDIX

Transcription Conventions

T: tutorW: writer(.) Each (.) represents a .5 second pause— speaker chaining/interruption; no pause between speakers// speaker overlap< > unintelligible( ) nonverbal actionital relevant portion of interaction is italicized[ ] voice modulation (e.g., loud, softly)

3

Negotiating an Institutional Identity:Individual Differences in NS and NNS Teacher

Directives

Lynda YatesLaTrobe University

In cross-cultural and interlanguage pragmatics research on requests there hasbeen a strong orientation toward understanding the commonalities in the wayspeakers from particular language backgrounds perform in different situations,that is, on intragroup similarities and intergroup differences in behavior (see,e.g., Blum-Kulka, 1982, 1985, 1987, 1989; Eslamirasekh, 1993; Fen, 1996;Fukushima, 1990; Ha, 1998; Huang, 1996; Kim, 1995; Koike, 1994; Kubota,1996; Lee-Wong, 1993; Mir, 1995; Rose, 1990; Trosborg, 1995; Yu, 1999;Zhang, 1995). This has left out of focus the issue of individual variation.Moreover, the assumption is usually made that the gap between any group ofnonnative speakers and the nativespeaker "control" or baseline data is that "theNS have an established pragmatic competence on which they can draw for thesituation" (Bardovi-Harlig & Hartford, 1996, p. 172), and that this is relativelystable and uniform. Such approaches therefore obscure what we all knowintuitively, as members of a speech community: that speakers vary in the waysin which they use language and project themselves as individuals in differentsituations. In this chapter, I draw on data collected in a school setting to explorethe issue of individual variation among both native speakers and Chineselanguage background speakers of English as they negotiate their institutionalidentities as trainee secondary teachers in Australian classrooms.

A focus on the commonalities shared by speakers of a speech communitycan provide important insights into what Gumperz (1996, p. 402) calls "specific,taken-for-granted, knowledge of background information and verbal forms" thatare shared by a community. However, this focus on the shared inevitably entailsa simplification of the way language is used and underwrites the construction ofbroad generalizations in which variation between individuals is neglected.Culture-specific styles and norms can only represent in the broadest outline thetrue diversity to be found within a culture, and we should be wary of regardingany language and cultural group as a "monolithic unity" in which individualdifferences are obscured. As Shea (1994, p. 380) warned: "Culture is not anessentialist construct where members adopt similar values, maintain uniformbeliefs, and have interpretive conventions."

Although cultural groups may differ in the way they behave and uselanguage in any particular social situation or identity, there will nevertheless be

67

68 Yates

differences between individuals in the way in which they behave linguistically(Hansen & Liu, 1997).

Social constructivist approaches to sociolinguistics have emphasized theactive role played by an individual in the construction of their social identity inany situation. Drawing on the communicative resources shared by thecommunity, speakers actively established their social identities, in part throughtheir use of language (e.g., Gumperz, 1982). They can do this by choosing toexpress certain acts in certain ways, or adopting a particular "stance" or"socially recognized point of view or attitude" (Ochs, 1993, p. 288). Theadoption of a particular stance may impact, for example, on how much certaintyto display and how to do it, or how much and what kind of emotion to display inwhat kind of situation and so on. Although the means for such displays are notdirectly encoded in a language, they rely on interpretation through convention,and the projection of social identities therefore crucially depends on theexistence of shared conventions, as well as a shared political and social historywhich associates the acts and stances used with the social identity intended(Ochs, 1993, p. 290).

However, although members of a speech community may draw on broadlyshared conventions for the display of social identity, each individual will drawon these differently, in line with the way in which they wish to presentthemselves (see Goffman's [1956] notion of demeanor). Differences betweenways in which individuals make use of these may relate to individualpsychology, personal expression, relationship with particular groups, or theimmediate interactive context (Gumperz, 1996, p. 376). Even within "the samegross social identity" (Ochs, 1993, p. 297), such as "parent," "student,""teacher," and so on, different speakers will construct their social identities indifferent ways. That is, although speakers may draw on the same conventions,speakers will use them variably to project different social identities, even withinthe same general social role, as they struggle to change what is normallyexpected from a social identity, or challenge their social identity over time.Moreover, individuals are not necessarily consistent across time or contexts; thesame speakers may project themselves differently on different occasions and indifferent situations. Social identities are dynamic, and individuals belong tomany different groupings in society; therefore, they have multiple identities orsubjectivities which vary across situations and time as they enact a variety ofroles (Weedon, 1987). People build multiple, compatible social identities thatmay be blended or even blurred. Thus, second language learners, like nativespeakers, find that "there are no simple social or linguistic formulae that spit outhow to compose suitable identities for the occasion" (Ochs, 1993, p. 298).

Because we are socialized into particular ways of acting in social situationsfrom an early age (e.g., Ochs, 1988, 1996), adults who arrive in a particularcontext in a new community may not be fully aware of what conventions,including what acts and what stances, are considered appropriate for thedifferent social roles enacted in that context. Thus the negotiation of a socialidentity in a context is especially challenging for those who are interacting in a

Negotiating an Institutional Identity 69

second language and operating in a social context with which they are familiarin general terms, but which is subtly different from that into which they weresocialized as children. Adult immigrants to Australia from the Peoples' Republicof China may not therefore be familiar with the shared conventions that nativespeakers (NS) can draw on in constructing their identities as teachers inAustralian English. They may have an incomplete awareness of the repertoire offorms that can be used to mitigate directives, and the exact force of these formsin interaction, or they may have incomplete or inaccurate sociopragmatic under-standings of how much and what kinds of mitigation are appropriate in thecontext, and what kind of stances are familiar to the students. That is, they maybe able to "perform particular acts and stances linguistically" (Ochs, 1993, p.291) but may lack awareness of how these relate to particular social identities ina second culture.

Moreover, an adult nonnative speaker (NNS) of any language has alreadyestablished identities which have found expression through interactions in adifferent language and culture. Although they may have spent many years inAustralia, and act in many ways like other members of the Australiancommunity, they may nevertheless hold values from their first culture (e.g.,Busher, 1997), and may not feel comfortable adopting stances and expressingsocial identities that are unfamiliar. As Hinkel (1996) shows, interlanguagespeakers may not accept aspects of the target culture, and thus may not want tobe completely nativelike in their language use. They may therefore seek todiverge from the ways in which native speakers express themselves as a meansof projecting their individual or group ethnic identity (e.g., Siegal, 1996).Indeed, if they are visibly ethnically different from the majority Anglo-background population, as most immigrants from the PRC are, they may not beratified in any role in the same way a NS might be: Native speakerlike behaviormay not be accepted from nonnative speakers, simply because they may lookand sound different from NSs (Amin, 1997; Tang, 1997). In this study, I wasoriginally motivated by the desire to see whether there were differences in theways that Chinese background and Anglo-background trainee teachers mitigatedtheir directives to students in the secondary classroom. This topic was ofpractical interest for me as a teacher-trainer, because of frequent reports ofmiscommunication difficulties experienced by Chinese background teachers inthe Australian school system. These are frequently blamed on languagedifficulties, but may actually relate to difficulty in establishing rapport withlearners and constructing a social identity as a teacher in an educationalenvironment built on very different, and less hierarchical, cultural traditions oflearning (see, e.g., Cortazzi & Jin, 1996; Hofstede, 1986, 1994; Scollon, 1999).Since directives are potentially face-threatening (Brown & Levinson, 1987),highly variable across cultures (e.g., Blum-Kulka, House, & Kasper, 1989;Huang, 1996; Yu, 1999), and reflective of social relationships, these werechosen as the focus of the analysis.

As the study progressed, however, I found that the rich, naturally occurringdata collected in classrooms allowed the examination of both the ways in which

70 Yates

language background appeared to influence the ways that the participants usedmitigation, and the differences in use of mitigation among individuals from thesame language background. This ability to examine both the differences and thecommonalities has illuminated some of the different stances that teachers take asthey attempt to negotiate their social identities as teachers in the classroom.

THE STUDY

The aim of the study was to investigate and compare the verbal mitigation ofdirectives among two groups of speakers: one which had grown up and beeneducated in Australia, and the other in the People's Republic of China. In doingthis, I wanted to explore the ways in which speakers pay symbolic attention toboth speakers' negative face needs of nonimposition and positive face needs ofapproval and belonging (Brown & Levinson, 1987) because, with few excep-tions (e.g., Scarcella & Brunak, 1981), most previous studies of interlanguagerequests have focused on the more formal devices of politeness, and have notdeveloped Brown and Levinson's essential insight. This neglect has been, inpart, a function of the type of elicited data usually analyzed in such studies,since data which has been elicited through the use of a discourse completiontask (DCT) or a role-play tends to be more formal and monitored (Hartford &Bardovi-Harlig, 1992; Lee-Wong, 1993; Rintell & Mitchell, 1989; Rose, 1992,1994). Speakers are likely to feel they are being required to put on aperformance for the researcher, and, where data are collected in written form,the very act of writing is likely to encourage more formal, "best" behavior,eliciting what speakers think they ought to do, rather than what they actually do.This favors the collection of devices associated with formal politeness, anddisadvantages the collection of devices associated with address to positive face,such as the signaling of common group membership through the use ofcolloquial expressions (which are seldom written).

The use of an institutional setting, therefore, offered a number ofadvantages. First, it allowed naturally occurring instances of language use in anauthentic context rather than the metapragmatic judgments or simulatedlanguage used collected through elicitation techniques. Second, it represented acompromise between ethnographic approaches in which the contributions of allkinds of speakers are collected, and more controlled elicitation methods inwhich speaker and situational variables can be more carefully controlled. Thusthe participants were all interacting in the same gross social identity in a similarcontext with broadly similar mid-term and long-term goals (to pass thepracticum and to teach in a secondary school). Although the type of lesson theytaught varied, there were similar phases to each class (setting up the activity,asking class questions, checking on individual work, directing practicalactivities, setting and returning homework, and so on). Moreover, a largenumber of directive tokens could be collected in ways that are replicable infuture studies, but which also allowed a close examination of function in their


interactive contexts. Each directive could therefore be examined for function ona number of levels.

Participants

For the study, 18 participants, nine Anglo-Australian background (AB) and nineChinese background from the PRC (CB) were audio- and video-recordedteaching two lessons in an Australian secondary school as they undertook thepractical teaching component of their initial teacher training qualification.1 Theywere all graduates studying at various universities in Melbourne. The CBparticipants were aged between 25 and 39, while the AB participants werenative speakers of Australian English, were slightly younger (between 21 and28), and had all been educated in Australia. The CB were recruited for the studythrough the help of teaching practicum placement coordinators in the educationdepartments of the major universities in Melbourne, who identified traineeteachers who met the criteria (i.e., who were born and raised in the PRC andwho were teaching a subject other than Chinese). I included in the study thewhole sample of trainee teachers who met these criteria and agreed to participateover a 2-year period. In general, most people I approached were sympathetic tothe aims of the study, and agreed to participate, although life events (in one casea car crash!) forced some to withdraw. I identified the AB participants throughmy position in the education department at a Melbourne university (La Trobe)

Table 3.1. Summary of characteristics of Australian background and Chinese background partici-pants

Characteristic

Age

Country ofEducation

Male

Female

Curriculumareas

Australian Background

Early to late 20s

Australia

5

4

BusinessESLMathMusicScienceSwimming

Chinese Background

Mid twenties to late thirties

People's Republic of China (1 part-educated inHong Kong, degree in Australia)

( l . y r l l )(2, young & yr 7)(2,yr7, 11)(l,yr7/8)(2,yr 1 1 & l l )(l .yrlO)

4

5

Business (1, yr 11)ESL (2, young & middle)Math (2, yr 8 & 9)Music (I,yr7)Science (2,yr lO & ll)Swimming ( 1 , yr 9)

Note: The school in which these sessions were recorded did not divide classes in ESL according tostrict year level but had a young (7/8) and a middle class (9/10).

1 In Australia, primary education runs from Kindergarten to Year 6, and secondary school from Year7 to 12. The students were therefore between 11-18 years old.

72 Yates

and was able to match the CB for subject area and gender as far as possible (seeTable 3.1).

Each participant was recorded in school teaching one of the two specialistsubjects they were training to teach upon graduation. These ranged fromswimming sports to math, and included business studies, science, and ESL, butnot foreign languages (see Table 3.1). The students in these classes generallyreflected the social and ethnic mix of the school community, except in ESLclasses, where all students came from a non-English-speaking background.Again, I enlisted the aid of practicum placement coordinators throughoutMelbourne in order to gain permission to record in schools. I needed permissionfrom all regional school authorities, the ethics committee of my university, theparticipants themselves, their supervising teachers, and the parents of all thechildren in every class. In the latter case, permission was obtained through asigned consent form distributed as soon as the participant arrived at the school.

For each of the two lessons recorded, the participants wore a small portablecassette recorder with a lapel microphone which was switched on at thebeginning of the class and left running until the end. Because most schools'scheduled classes to run for approximately 50 minutes, this enabled the majorityof the lesson to be captured on a 45-minute tape. I also video-recorded eachlesson and collected field notes so that the context of any directive could beclearly identified. After the recordings, I interviewed each participant about theirexperiences with the practicum. The classroom discourse recorded in this waywas transcribed, and any directive given by the teacher identified.

Directives: Identification and Coding

Using Searle (1975) as a starting point, directives were broadly defined asattempts by the teacher to get a student to do something concrete in the future.No attempt was made to distinguish directive acts of different force (e.g., anorder from a request), but certain kinds of pedagogical acts, such as generalsolicits, were not included in the analysis. Each directive was then coded foraspects of the context (e.g., whether they were made to a single person or thewhole class) and for the mitigation devices used to soften it. Mitigation wascoded in four major categories identified by Blum-Kulka et al. (1989): thestrategy (level of directness) of the head act of the directive; syntacticmodification of the strategy; lexical modification of the directive, andpropositional modification (the use of additional moves to support the head act,called external mitigation in Blum-Kulka et al., 1989). The framework wasexpanded to include other devices identified in the data, and to include attentionto both negative face concerns of nonimposition and positive face concerns forapproval and belonging (Brown & Levinson, 1987; O'Driscoll, 1996).

To capture the ways in which positive face needs were addressed inmitigation, the coding framework included some specific devices which signaledwarmth and approval (such as the use of praise or approval propositional movesto support the directive), or solidarity through the reduction of social distance


(such as the use of humor or inclusive cultural references). The reduction ofsocial distance was also tracked by assessing the extent to which the particularforms chosen by the participants reduced, maintained, or increased socialdistance. Thus more formal devices associated with formal politeness (such ascould you and please) were considered to be more distancing than morecolloquial forms such as the use of a solidary address term, such as guys or avernacular term, as in the example in Table 3.2. A summary of the codingframework can be found in Appendices A and B, and examples of some of theways in which social distance was reflected in the use of different devices isgiven in Table 3.2

In order to allow comparison with previous studies and some quantitativeexploration of the data, the use of each category of mitigation and tokens ofparticular devices were counted and expressed as a percentage of the totalnumber of directives used by an individual or a group. That is, the frequencywith which particular types of mitigation were used was calculated both forlanguage and gender groups (i.e., for AB females and males, and for CB femalesand males), as well as for individuals. This allowed the identification of bothmain trends in the data, and individual patterns of variation. The data were alsoexplored qualitatively using the NUD*IST2 program.

COMMONALITIES AND DIFFERENCES IN MITIGATING STYLE

As discussed earlier, an important overall finding of the study was that, althoughlanguage background was important in explaining how the participantsmitigated their directives, there were also considerable individual differencesthat appeared to relate, not only to gender and level of pragmatic sophisticationin English, but also to differences in individual mitigating style, which reflectedthe way in which participants chose to project themselves as individuals andteachers in the classroom. These are discussed first in the next section, beforediscussion of the ways in which these styles differed and the extent to whichthey can be related to background discipline, life experience, and other factors.

Overall Trends for Language Background and Gender

The overall trends for both males and females from the two languagebackgrounds can be seen in Table 3.3, which presents the number of mitigatedand unmitigated directives produced by each group. Because the number ofparticipants in each language and gender grouping is small (three to five), grouptotals should only be considered a rough guide to patterns of use and interpretedin conjunction with the results for individuals discussed in the next section.Nevertheless, the frequencies shown in Table 3.3 suggest that both languagebackground and gender influenced the mitigation of directives, since AB miti-

2 An ethnographic data processing program developed by QSR.

74 Yates

Table 3.2. Examples of address to different types of face

Strategy

Distance-maintaining e.g., you +stem or you will:• you copy it down• you are here. / you will start

from here

Less distancing, e.g. 'going to' or 'Iwant':• alright /so first we 're going to

listen now/ 1 want the boys allto move/ thata way/ ok

Lexical mitigation

More distancing, e.g., politenessmarker• if you still talk you sit in there

please, OK?

Less distancing, e.g., vernacular:• grab a copy of that

Syntactic mitigation

More distancing: e.g. past tense:• would you go and speak to Mr. X please

(name) ?

Less distancing, e.g., going to and defocalisers:• you are going to have to stop distracting all

the people around you• so, what 1 suggest you do/ is just read

through it

Propositional mitigation

Warm but not solidary, e.g., approval moves, e.g.,Praise• ok that was good /come in here guys

Less distancing solidary rapport moves, e.g.,personal/cultural reference:• get off it (name) it's not a surfboard alright?

gated more frequently than CB, and, within each language group, femalesmitigated more than males. The overall pattern of mitigation use follows thepattern AB females (72.7%) > AB males (64.8%) > CB females (58.2%) > CBmales (54.8%).

The different groups also favored the use of different categories ofmitigators. Table 3.4 shows the frequencies with which the groups of partici-pants used the four categories of mitigation. In this table, the raw numbers ofdirectives that contained a particular category of mitigation are reported as apercentage of the total number of directives given by that group. Reading acrossthe second row, for example, we see that the four female AB participantsproduced a total of 631 directives, 208 or 33% of which used an indirectstrategy. As the total number of directives produced varied widely betweengroups and individuals, percentages were used as the basis of comparison.

It is important to note that the categories of indirect strategies and syntactic,prepositional, and lexical mitigation (Tables 3.3 and 3.4) are not mutuallyexclusive. Any one directive can combine an indirect strategy and more than onemitigator. Reading across Table 3.4 in the first row we find that among the totalnumber of directives produced by AB males, 24.5% exhibited indirect strategies,9.7% used syntactic mitigation, 22.5% prepositional mitigation, and 38.3% usedlexical mitigation. As Table 3.3 shows, 35.2% of the directives had nomitigators. This shows the extent to which directives occur with multiplemitigators. In other words, some directives have no mitigators whereas otherdirectives have more than one type.


Thus AB of both genders were less direct than CB males, and both ABgroups used more syntactic and prepositional mitigation than either CB group.Although there was less difference between CB and AB in the frequency withwhich the groups used lexical mitigation, as discussed in the following section,many CB tended to rely on the use of minimal tags or token agreements, whileAB used a wider range of both minimizing and solidarity devices. Female ABtended to use most devices more frequently than their male counterparts. Thiswas not always the case among CB, where level of pragmatic sophistication inEnglish appeared to be a factor (see the section entitled Variation Among theChinese Background Speakers).

Whereas overall comparisons with previous studies of Chinese backgroundinterlanguage speakers of English (Fen, 1996; Yu, 1999) are problematicbecause of differences in data collection and analysis (see Yates, 2000), theseresults nevertheless parallel those found for learners in Fen (1996) (the onlystudy that investigated interlanguage speakers in high-power positions), andshow some of the tendencies outlined for Chinese speakers of Mandarinsuggested by Lee-Wong (1993). That is, Chinese background speakers inpositions of relative authority mitigate less than do native speakers of English insimilar roles: They are more direct, and use syntactic and prepositionalmitigation less frequently than native speakers of English, and rely more heavilyon lexical mitigation.

Individual Variation

However, there was also considerable individual variation in the participants'use of mitigation which does not entirely fit within these patterns. Figure 3.1shows individual participants ranked according to the frequency with which theyused mitigation. Exact frequencies are shown in Appendix C.

Table 3.3. Mitigation of Directives by Language and Gender

Language/gender

AB m

AB f

Tot AB

CB m

CB f

Tot CB

Total indirect directivesand/or directives with

mitigation

n

328

459

787

241

399

640

% of tot

64.8

72.7

69.2

54.8

58.2

56.8

Total directives withoutindirectness or mitigation

n

178

172

305

199

287

486

% of tot

35.2

27.3

30. 8

45.2

41.8

43.2

Totaldirectives

n

506

631

1137

440

686

1126

76 Yates

Table 3.4. Mitigation of Directives by Language and Gender

Language/ Total indirectgender strategies

AB m

AB f

Tot AB

CB m

CB f

Tot CB

n

124208

332

79

134

213

% of tot

24.533.0

29.2

18.0

19.5

18.9

Directives withsyntactic

mitigation

n

4998

147

24

24

48

% of tot

9.715.5

12.9

5.5

3.5

4.3

Directives withpropositional

mitigation

n

114

173

257

65

120

185

% of tot

22.527.4

25.2

14.8

17.5

16.4

Directives withlexical mitigation

n

194295

489

173

253

426

% of tot

38.346.8

43.0

39.3

36.9

37.8

Totaldirectives

n

506631

1137

440

686

1126

Each participant is identified by two initial letters that indicate whether theyare Anglo-Australian (A) or Chinese background (C), female (f), or male (m),followed by a number that identifies them uniquely (as in Afl or Cml). Amitigation index was calculated for each individual participant by addingtogether the number of directives that contained lexical mitigation, prepositionalmitigation, syntactic mitigation, and indirect strategies and dividing the sum bythe total number of directives to give an overall mitigation index. Because anindividual directive can contain a mitigator from more than one category ofmitigation (and possibly from all four), the total index may amount to more than1.0. The results for individuals are presented in the order of this ranking, fromthe participant who had the highest mitigation index, Afl with a total of 1.69(mitigator categories per directive) to the lowest mitigator, Cm4 with an index ofonly .42 (fewer than one mitigator category for every two directives). Fromthese results, it can be seen that individuals varied considerably in the frequencywith which they used mitigated directives in general, and also in their use of thedifferent categories of mitigation. Thus, for example, although participants Cf3and Am3 had similar mitigation indices overall (.86 and .85 respectively), theytended to use different categories more frequently: Cf3 tended to use moreindirect strategies and less syntactic mitigation, whereas Am3 used lexicalmitigation, particularly vernacular, more frequently, but was generally moredirect in the strategy he used in the head act.

From Figure 3.1 it is clear that, although language background and genderwere important predictors of how frequently participants used mitigation withtheir directives, individuals varied in relation to the general tendencies for theirlanguage and gender peers. Thus AB females, as might be anticipated from theoverall results presented in the previous section, generally mitigated morefrequently than the other participants. They provided the two highest users ofmitigation, Afl, and Aft, who had mitigation indices of 1.69 and 1.26respectively. However, two of the four AB females, Af3 and Af4, mitigated less


Figure 3.1. Individual Variation in Frequency of Mitigation

frequently, and had somewhat lower indices of 1.14 and .92, respectively. ABmales generally mitigated slightly less frequently than the females, with indicesbetween 1.17 and .77. Although, in accordance with the overall trends discussedin the previous section, CB participants in general mitigated infrequently (five ofthem exhibiting low indices ranging from only .63 to .42), two CB participants,Cf1 and Cml, nevertheless ranked third and fourth overall in the frequency withwhich they used mitigation, with mitigation indices of 1.22, and 1.19respectively.

It is clear, therefore, that although background and gender were important inunderstanding how frequently a participant mitigated, individuals within thesame language and gender group varied considerably. This variation was alsoevident in the types of devices that participants used, as becomes more evidentin the following section where individual mitigating style is discussed.

Mitigating Style

Both AB and CB trainee teachers varied considerably, not only in howfrequently each used a particular category of mitigation or device, but also in thetype of device they chose to use and the "teacher" social identity they wereprojecting through their choices. However, in one respect all CB were similar

78 Yates

Table 3.5. Individual Styles and Factors

Afl,2,3

Ami

Am2

Am5

Af4

Am3

Am4

Cfl

Cml

Cf2

CO

Cf4Cf5Cm2Cm3Cm4

Style

Friendly,solidary, non-impositive

Matey. Academic

Matey, non-impositive

Super-matey,'sports '

Authoritative, butnon-impositive.Mildly solidary.Signalsseparateness.

Vacillating,authoritative

Warm, friendlybut non-solidary,non-impositive.Emphasisesseparateness.

Warm, non-impositive,separate, a littlesolidary

Warm, distantand limited

Impositive andvery distant,sometimes warm

Features

Low-distance, highlymitigated style

Medium mitigation stylewith low-medium distance

Medium mitigation, but lowdistance. Lower lexical

Low frequency ofmitigation, plus lowdistance. Relatively direct,high

Medium mitigation stylewith medium distance.Relatively indirect, highsyntax

As for Af4, but overall lessindirectness and syntactic,more propositional andlexical. Malecharacteristics.

Vacillation betweendistanced and solidary

Highly-mitigated, butminimization more target-like than address to socialdistance

Highly-mitigated,minimisation more target-like than address to socialdistance, but used somemore non-distancingdevices than other CB

Indirect and relativelyfrequent syntax comparedto other CB, but littleflexibility or solidarity

Non-target-like address toall dimensions. Frequentfeatures indicating languageproficiency difficulties

Factors

Female, young, projecting asolidary identity

Male-influenced discipline,(engineering), young, solidary

Young, less gendered discipline(drama)

Solidary, male-oriented discipline(sport)

Older/experienced, content-focused, tertiary-oriented. Female

Confident, tertiary-oriented. Male.

Young, confused about teachingidentity

Pragmatic sophistication inEnglish, inexperience of role,desire to fit in. Sought outEnglish-speaking room-mates.

Pragmatic sophistication inaddressing negative face needs,desire to fit in, some experience ofrole. Sought out English-speakingroom-mates

Partial pragmatic sophistication.Less intense experience of nativespeakers.

Lower pragmatic sophistication inEnglish, limited understanding ofteacher role in Australia, lessintense experience of nativespeakers, language proficiencyissues


and differed from AB: They tended to use a mitigating style that was lesssolidary. That is, AB participants regularly addressed positive face needs ofbelonging and inclusion, by choosing forms that reduced, or at least did notincrease, social distance. None of the CB used such devices to the same degreealthough they used formal politeness markers such as please more frequently.They did not, therefore, seem to reduce social distance to adopt the more"matey" stance taken by the AB teachers. This can be seen in Table 3.5 whichsummarizes the different styles used by the participants. In addition, among theCB, awareness of the types of usage of mitigating devices appeared to be afactor. The mitigating styles adopted by the AB and the CB participants areexplored in greater detail in the following two sections.

VARIATION AMONG THE AUSTRALIAN BACKGROUNDPARTICIPANTS

In addition to the considerable differences between AB individuals in thefrequency with which they used the categories of mitigation, there were alsodifferences in the forms and devices they used, and in particular in the degree ofsocial distance they projected vis-a-vis their interlocutors through their choice ofmitigating device. These differences seem to have been motivated, by gender,discipline background, and context, and also by other, more individual factorsrelated to the way in which the participants each negotiated their identities asteachers. Those who mitigated more frequently tended also to use distance-reducing devices more frequently, and distance-maintaining devices lessfrequently, than those AB participants who mitigated less, as can be seen bytheir more frequent use of all categories of mitigation. Table 3.6 summarizes theuse of the four categories of mitigation by the Australian backgroundparticipants, together with an assessment of the degree of distance they projectedbased on the forms and types of devices they used.

As shown in Table 3.6, those AB who used the most mitigation, Af2 Af1,Af3, and Am1, also tended to adopt a lower distance approach to mitigation thandid Af4, Am4, and Am3. Males Am2 and Am5 were exceptions to this tendency.This difference in the projection of social distance could be seen in all fourcategories of mitigation, that is, in their choice of strategy, in their syntacticmodifications, lexical choice, and the prepositional support they gave theirdirectives, although some participants tended to be more distanced in theirapproach in one category of mitigation than in others. Af1, Aft, and Am 1 wereconsistently low distance in their approach to all categories. Af3 adopted a lowerdistance approach in her use of syntactic, prepositional and lexical mitigation,but not in her use of substrategies. Context may have been important here, sinceshe was teaching a lively class so that a relatively high percentage of herdirectives related to discipline issues in response to perceived misbehavior. Thisseems to have encouraged her to adopt greater distance in her choice of strategy,as in the following examples. In Example (1), she chooses a more distancingconventionally polite formula 'would you' to increase social distance as she

80 Yates

Table 3.6. Overall Use of the Categories of Mitigation by the Australian Background Participants

Participants

AmiAm3

Am4

Am5

Am2

Aft

AflAf3

Af4

Indirectnessstrategies

VV-+

+++

V

Syntacticmitigation

V++-V

+++

+

Prepositionalmitigation

+V—V+

++

Lexicalmitigation

+

—+-++

+

-

Distance

LowHigh

High

Sports

Low

LowLow

Low

High

Note. + = frequent use; = moderate use; — = low use (relative to all 18 participants, including CB).

asserts her authority, while in Example (2), she reinforces the formal politenessof 'would you' with the formal and distancing on-record marker 'please':

(1) (Af3)

Would you boys stop being silly because you were working so well.

(2) (Af3)

Would you go and speak to Mr. Smith3 please Ryan?

Thus three of the four AB females (Afl, Af2, Af3), used mitigationfrequently and reduced social distance through their choices of mitigation. Theydid this by using fewer distancing strategies, modifying their directives with lowdistance syntactic mitigation, such as deflectors, as in Example (3), or heavy useof positive politeness hedges as in Example (4) taken from a business class withYear 11 students. All three were relatively young (early 20s) and seemed to beprojecting a friendly, solidary persona, in which directives were softened byappeal to in-group membership and reduced social distance.

(3) Solidary female (Af3)

So, what I suggest you do/ is just read through it,

3 All names have been changed.


(4) Solidary female (Afl)

OK. /Now like beside that /I'm sorry I've (there's not) not much room on the board/you can put something like/ 'if the store has a run down look and stock looks likeshop worn '/[..] that sort of thing

Female participant Af4 used mitigation with her directives slightly lessfrequently than her female peers, as she used fewer lexical devices (e.g., just or /think) and maintained more social distance than the other three. Unlike the otherthree, for example, she did not use a single positive politeness hedge (e.g., like,sort of), and used vernacular and personalization (e.g., for me) less frequently.She did, however, make relatively high use of syntactic mitigation using modalsand conditionals, which tended to be a little more distancing, as in Example (5)and Example (6), and used more distancing lexical mitigation, such as pleasemore frequently, as in Example (7).

(5) Authoritative female (Af4)

Year 11, would you like to take a five minute break now?


Now, if you can get into groups of two,


Can you two share a text book between you, please

Her rather more formal mitigating style may be, in part, related to regionaldifferences in that Af4 had grown up in a country town, whereas the other threehad grown up in Melbourne. It may also reflect the fact that she had hadexperience of tertiary teaching, while the other three had not, and that she wasalso a little older. Although the exact influences on her mitigating style remainunclear, it seems that she positioned herself in a more authoritative position inrelation to her students through her less solidary approach to mitigation.

Individual variation was even more evident among the AB males, and thismay reflect the influence of changing gender roles and perceptions of manhoodas Anglo-Australian culture absorbs the impact of feminism and the renegoti-ation of gender roles in the postindustrial era. While all AB males mitigatedtheir directives less than the three high-mitigating females, as shown in Table3.6, two males, Am1 and Am2, adopted a relatively high-mitigating and low-distance approach to mitigation. This low distance style projected a mateyrapport through frequent use of prepositional mitigation, and the reduction ofsocial distance through choice of strategy and lexis. The directive in Example(8) given by Am1 starts with "yeah," and so signals alignment. It also projectsauthority, in that he clearly states what he wants done by using a direct "I want"strategy, but this is modified syntactically to become the softer "I'd like." Headds a reason, so that the students know why he is directing them in this way,

82 Yates

and addresses them using the vernacular "guys" to emphasize solidarity ratherthan social distance.

(8) Matey, academic (Am1)

Yeah. I'd like it in your own words 'cause that way I can see whether you guysknow what you're doing.

The style used by Am2 was less direct than that used by Ami (as in Example9), and he used lexical mitigation much less frequently. He seemed to beadopting a style that was masculine and matey, but also relatively nonimpositiveand inclusive. Although they both used support moves frequently to justify orcontexrualize their directives, Am1 tended to provide reasons, whereas Am2tended to use preparators and sequence moves (which let students know whatwas happening next, as in Example 10). In addition, like the three high-mitigating females, he used positive politeness hedges and personalization in anapparent attempt to signal solidarity.

(9) Matey, nonimpositive (Am2)

George, do you want to give me some homework on Monday?

(10) Matey, nonimpositive (Am2)

and if you want to finish your report for homework 10 and give it to me, /'// gothrough it and correct it for you OK?

Background discipline and context may have influenced the type ofclassroom persona that each wanted to project. Participant Ami trained inengineering, a traditionally male-dominated profession, in which all-male dis-course is common. Inasmuch as interlocutor gender is an important influence onthe use of politeness phenomena (Holmes, 1995), it is likely that norms ofinteraction within this profession are more traditionally masculine, and this mayunderpin Ami's use of directness and solidarity. Participant Am2, on the otherhand, trained in media studies, which is a more "feminized" discipline,traditionally more tolerant of different ways of expressing gender, and this mayhave influenced his tendency to use indirect strategies. His lower use ofvernacular, however, may also have been a reflection of the modifications madeby teachers who are teaching learners from other language backgrounds (i.e.,"foreigner talk"; Ferguson, 1975).

Two other male participants, Am3 and Am4, adopted a more distancedapproach and used mitigation less frequently. These two not only mitigatedoverall less frequently (see Figure 3.1), but also made relatively high use ofvarious distance-maintaining devices (e.g., please and past tense syntacticmitigation) and relatively low use of distance-reducing devices (such as positivepoliteness hedges and prepositional mitigation). Although both appeared to be


projecting distance in order to assert their authority, they seem to have done thisfor different reasons. Participant Am4 seems to have projected distancereactively, in response to discipline difficulties. Unlike Af3, however, his irrita-tion showed and was noticed by the students. As is evident from the classroomtranscripts of his lessons and from his own reports in the postrecordinginterviews, he was not always in control of the class and seemed to waver in hisapproach to the students, at times appearing to want to establish a good rapportwith them and at others appearing to emphasise the power differences in anaggressive way. At one point, one of the students inquires of him, "You don'tlike this class, do you, Sir?" In Example (11), he uses a conventionally politeform ("could you") to request a book for some students who were not well-prepared for the lesson. He also refers to the students for whom the book isdestined as "clowns," thus apparently colluding with the girls and signalingseparateness from the boys.

(11) Vacillating (Am4)

Could one of you girls lend these two clowns in the front row a book? They'lllook after it. Thanks

On occasion, he switched from a more direct style to a more conventionallypolite and distancing style in response to noncompliance. In Example (12), he isattempting to get his students to come into the classroom. He starts with a directhead act, reinforces this with an elliptical directive and finally resorts to formal,conventional politeness as he signals his displeasure at their unsatisfactorybehavior and makes them go back outside to repeat the requested action.

(12) Repetition (Am4)

So come in quietly.Quietly Year seven.Excuse me Year seven could you go back outside and line up please.

Both Am4 and Am3 used high amounts of vernacular, however, whichreflects the fact that they used both formality and informality as part of amitigating style that both distanced and signaled solidarity. Participant Am3,however, used the projection of distance proactively. In a similar way to Af4, heappeared to project a certain amount of distance, deliberately in order tomaintain authority (i.e., in the absence of any discipline difficulties). In Example(13), Am3 uses an Australian masculine diminutive ending in -o, thus signalingsolidarity, with a conventionally polite request with please.

(13) (Am3)

Danno, can you put that chair down next to you please?

84 Yates

The fifth AB male, Am5, who was recorded teaching swimming classes,adopted a style that was very direct and very matey (Example 14). His use ofdevices of nonimposition (e.g., indirect strategies) and distancing syntacticmitigation was very low, but he made frequent use of lexical mitigation, particu-larly vernacular, and of propositional mitigation, including humor and personal/cultural references, which were sometimes bantering and apparently abusive. Hethus drew heavily on solidarity devices and low-distance forms. This style seemsto have drawn on the discourse of sporting events, particularly male sportingevents in Australia. Kuiper (1991) noted similar tendencies in sporting events inNew Zealand, and Deby (2003) in ice hockey commentaries in Canada. Deby(2003) suggests that such language use relates to displays of the ability to bothdole out and withstand psychological as well as physical intimidation, a sort ofverbal correlate of actual sporting prowess. The fact that Am5 was recorded in aswimming class in an all-male school seems to have encouraged the use of thiskind of particularly "masculine" style of mitigation. Note the use of a directstrategy and the solidary term of address, "mate" in Example (14).

(14) (AmS)

Get out of the pool mate,/yom nose is all bloody

Thus, although there were commonalities in the way AB mitigated theirdirectives, each drew on these in often subtly different ways to project theirsocial identity. Each was constructing himself or herself as a male or female, asa teacher and as an individual in the classroom context where their previousexperience was as a student rather than as the authority figure, and each did thisdifferently in relation to what may be considered customary in Australia for theirgender. Such insights remind us that variables such as language background andgender are influential but not deterministic. Rather, they should be seen asdynamic influences which are themselves in a state of constant change. Similarvariation might also therefore be expected among the CB. However, as notedearlier, the picture here is further complicated by issues related to grammatical,pragmalinguistic and sociopragmatic proficiency in English and in familiaritywith the gross social identity of being a teacher in an Australian school.

VARIATION AMONG THE CHINESE BACKGROUND PARTICIPANTS

Like the AB participants, the ways in which the CB mitigated their directiveswere also influenced by the type of identity they wished to project in theclassroom. However, they not only had been socialized into a culture drawing ondifferent views of personhood, gender, and social relations, but were alsoconstrained in the ways in which they could relate to target culture norms bytheir linguistic and pragmalinguistic competence in English. Several of themseemed to have a less well-developed repertoire of mitigating devices at theirdisposal, as well as a potentially different understanding of the force these may


have in an Australian classroom. Thus one factor that seems to be important inunderstanding is the variation among the CB is their level of pragmaticsophistication in English, which I characterize here in terms of the size of theirrepertoire of mitigating devices and the way in which it was used.

As shown in Figure 3.1 and Table 3.7, the CB varied widely in howfrequently they used mitigation, from the three highest users of mitigation, Cf1,Cml, and Cf2 (with mitigation indices of 1.22, 1.19 and 1.02, respectively), tothe low use of mitigation by Cm2, Cf4, Cf5, Cm4, and Cm3, with indices ofbetween 62.8 and 41.7. Participant Cf3 seemed to fall between the two groups.The former group appeared to be more pragmatically sophisticated in Englishthan their peers. They mitigated more frequently, used more indirect strategies,propositional support moves, and syntactic and lexical mitigation than theirpeers. They offered more options (through, e.g., using an interrogative ratherthan an imperative form) and used devices which minimised the imposition ofthe request (by, e.g., downtoning or underplaying the request with devices suchas just or for a minute). Their patterns of mitigation converged more closely onthose of the AB. However, while they did reduce social distance more than theirCB peers, they did not do this as frequently as did AB. These trends wereevident in their use of all four categories of mitigation. As shown in Figure 3.1and Table 3.7, however, they also differed from each other in some respects intheir mitigating style.

As shown in Figure 3.1, Cm 1 mitigated a much higher percentage of hisrequests than any other CB male. In this, he was closer to the general pattern forthe younger AB females than to that for the AB males. Although he mitigatedhis directives with roughly the same frequency as Ami, his style differed in thathe used indirect strategies far more frequently (ranking second in his use of

Table 3.7. Overall Use of the Categories of Mitigation by the Chinese Background Participants

Participants

Cm1.Cm3

Cm2

Cm4

Cf2Cfl

Cf5

Cf4

Cf3

Indirectness

+—

—

-

V+

-V

Syntacticmitigation

+

—

—-

VV

-V

Propositionalmitigation

+—

—

-

V+

-V

Lexicalmitigation

V—VV+

VV--

Solidarity

V——-

VV

--

Note. + = frequent use; = moderate use; — = low use (relative to all eighteen participants, includingCB).

86 Yates

indirectness compared to 12th for Am1). This was because he made frequent useof the relatively distancing conventionally indirect strategies, "can you" andfewer of the directives phrased as "I want" or "you have to" which were favoredby Ami. He did not use the high levels of vernacular and personalizationcharacteristic of Am1 (Example 15), but used distancing forms such as "please"frequently (whereas Am1 did not use "please" once!) and addressed his studentas "kids" as in Example (16), rather than the more solidary form "guys" used byAm1. Unlike Am1, he was therefore quite distancing in his mitigating style.

(15) (Am1)

I'd do it with that one I reckon/ 'cause some of these calorimeters I think are a bitstuffed4

(16) (Cml)

That's too much noise, kids, keep it down, please

He had been in Australia for 6 years, and seems to have made a veryconcerted effort to gain as much exposure to Australians and English usage aspossible. He had deliberately moved into shared accommodation withAustralians, and had worked in the retail industry where he used English daily.From his general use of language, these strategies seem to have paid off in thathe was largely accurate and commanded a wide range of devices. Although heseems to have understood enough about interpersonal pragmatics in theAustralian community to realize that requests, even to those in lower socialpositions, are frequently mitigated, he perhaps did not have sufficient familiaritywith how to do this as a male in a position of authority. Whereas Ami mitigatedfrequently, he was able to establish his authority clearly through the use ofrelatively direct directives, but mitigate their impact through a reduction ofsocial distance, and this seemed to be important in the way that at least some ofthe AB males projected their identity as teachers. Perhaps because he was notquite able to establish his credentials as a teacher, Cml was the only participantwho failed the teaching practicum on which he was recorded. Moreover, it wasonly when he viewed the video recordings of his classes that he understood whyhe had failed!5

Female Cf1, like Cml, mitigated very frequently, was relatively indirect inher strategy use and supported her directives with additional propositions, as inExample (17). She used lexical devices less frequently, although, like the otherhigh-mitigating female CB, Cf2, and the AB females, she used hedges morefrequently. She was often less accurate in her use of grammar, as can be seen inExample (18). Her overall style was warm and friendly rather than solidary,

4 Which could be glossed as "Use that one because some of these calorimeters are not workingproperly."5 He subsequently undertook a further practicum placement which he passed.


which is to say she communicated warmth and approval, but she did not usedisplays which actively reduced social distance to signal solidarity. For example,like Cf2, she used very little vernacular and no personalization at all. She used arestricted range of devices, albeit with a frequency that was often similar to thatof her AB counterparts.

(17) (Cf1)

Rosie, did see your homework I correct for you? I just gave to you a correctedhomework, could you correct them? And then I can tell you understand

(18) (Cf1)

or you maybe you are not all the same problem you want to work indiv individu-ally and you can

Overall, female Cf2 was far more accurate in her use of English generally,and she used a wider range of mitigating devices. Her style was also a little lessdistancing than that of Cml and Cfl, as she very occasionally used lessdistancing syntactic devices (such as the future as in "you're gonna do....")which no other CB did. Like AB in general, when she used a direct strategy itwas often in the form of "I want" rather than an imperative. Both Cfl and Cf2had lived with and had partners who were native speakers of English, and hadworked with young native speakers, Cfl as a nanny, and Cfl as a technicalassistant in a school. It seems that these experiences had enabled them to have acloser appreciation of how directives are mitigated, although Cf2's greateroverall linguistic proficiency allowed her to use a wider range of devices thanCf1.

Participants Cf5, Cf4, Cm3, Cm2, and Cm4 appeared to be pragmaticallyless sophisticated in English, since they relied more heavily on minimal lexicaldevices such as please, the downtoner just, agreement (e.g., yeah) and tokentags. They tended to use direct strategies, particularly the imperative, as inExamples (19) and (20) and "you assert" (you plus a stem form), as in Example(21), more frequently. Where they did use indirect strategies, these tended to beformulaic conventional requests (e.g., "can you") rather than the apparentlyadvisory strategy (e.g., "you can" or "you should") that AB participants usedmore frequently. They made very little use of syntactic mitigation, and whenthey did it was mostly in formulaic uses of the past tense, and their use ofprepositional mitigation was also generally lower. They, therefore, exhibitedmitigation behavior that more closely reflected that described in the literature onthe realization of directives in Chinese (Fen, 1996; Huang, 1996; Lee-Wong,1993, 1994; Yu, 1999). These findings suggest that they may have transferredfrom their first language a tendency to rely on economical lexical means ofmitigation (Lee-Wong, 1993) or hierarchical understandings of the teacher-student relationship which licence the routine use of bald directives to students.

88 Yates

(19) (Cf5)

Take a your note book out.

(20) (Cm4)

Ah boys boys () take em (xx)

(21) (Cf4)

it your homework, you exercise now.

However, there were still some observable differences in their styles.Participant Cm3 appeared to be very limited by his linguistic proficiency,perhaps because he had only been in Australia for 3 years. He used only alimited range of mitigating devices and relied on a few more salient formulaicpoliteness formulae which tended to be more distancing, such as could you(Example 22). This type of address to negative face is all the more distancingbecause it takes place in a male sporting environment. As we saw earlier, his ABcounterpart (Am5) who was also recorded teaching a water sports class, had astyle which was direct like that of Cm3, but highly solidary. Am5 used mostlydirect strategies, but they were mitigated with vernacular, inclusive addressterms, personal references, and even swearing, as in Example (23).

(22) (Cm3)

Now first training because the here not in deep enough () could you please moveto the deep side (students swim to deep end)

(23) (Am5)

T: What are you doing?St: xxT: Don't, Alan, that is bloody stupid, don't do it. Alright? Particularly ifsomeone's coming down. () You of all people should know better

Male Cm2 used very many direct strategies, and displayed very little graspof syntactic mitigation. It appears from other aspects of his teaching discoursenot analyzed here that he was trying to project a warm teacherly identity. Hisheavy use of "please" was somewhat distancing, although it illustrates someunderstanding of the need to mitigate requests to students. Females Cf4 and Cf5also projected a distanced identity. Their directives were often short and direct.Participant Cf4 often projected irritation and impatience, and appeared intolerantof the off-task activities of the young learners in her class, which she sometimesattempted to deal with by using directives with threats, as in Example (24). It isinteresting to note, too, that this extract shows that she was able to manipulate aconditional construction, although she never used it as a mitigating device, as


did AB, as in Example (25). This suggests she was either unaware of thepragmalinguistic function of the conditional, or of its relatively frequent use indirectives by teachers.

(24) (Cf4)

(XX taps watch) If you don't want to go home .. . if you don't want to exercise,if you keep talking you see me after lunch, ok, (=start now)

(25) (Af4)

Now, if you turn to page 147 in the text,

In contrast to the apparently irritated style of Cf4, Cf5's somewhat baldmanner of delivering directives (as in Example 19, "Take your notebook out")seemed to be more closely associated with a deliberate desire to project ateacherly identity from a position of superiority and distance. Support for thisview came from discourse overheard after the recording devices had beenswitched off. It was the last class of her teaching practicum, and once she wasassured that she had "passed" she seemed to relax and engaged in socialexchanges with the students sitting in the front rows.

Thus the Chinese background teachers differed in the use they made ofmitigation for a variety of reasons. In part, this variation may have been due tothe fact that they were at different levels of language proficiency, and this mayhave influenced their ability to use certain devices. In addition, though, they alsoseemed to be at different levels of awareness of how classroom directives mightbe formulated and of the impact that different directive styles might have oninterlocutors. Although the focus of this analysis was on the teachers' use oflanguage rather than on students' responses, there were some indications in thedata that the students reacted unfavorably to some aspects of mitigating styleused by CB. Thus, after a rather abrupt directive by Cf5 ("notebook out"), onefemale student uttered under her breath "Yes, Sir!", although Cf5 either did notnotice the comment or did not understand its force. Similarly, Cf3 farewelled herclass using the address term "boys and girls". After the recording equipment hadbeen turned off, one of the students in the class said to another sotto voce,"We're not boys and girls, we're guys." This seemed to indicate some resent-ment among the students that they were being constructed as children rather thanas apparent equals. Again, Cf3 appeared unaware of the impact of her use of thisaddress term.

MAKING SENSE OF DIVERSITY

As Ochs( 1993) noted,

communities often differ in which acts and stances are preferred and prevalentcultural resources for building particular identities. In one community, a stance

90 Yates

or act may be widely used to construct some social identity, whereas in anotherthat stance or act is rarely drawn on to construct that identity, (p. 300)

Thus cross-cultural comparisons may be usefully made of the conventionsrelating to how different acts and stances are performed and identified, and howthey are related to particular social identities. The overall trends in mitigationhighlighted in this study suggest that the adoption of a solidary stance might bean important aspect of some teachers' social identity in the classroom, and thatthis may be either unfamiliar or difficult for those raised and educated in aChinese learning culture.

However, as I have shown, individuals vary in the way in which theyproject themselves through language use, in this case, in how they mitigatedirectives in the social identity of (trainee) teacher. The variation found here is auseful reminder that all speakers of a language, both native and normative, areactively engaged in projecting themselves as individuals in relation to culturalnorms with such resources as they have at their disposal. Although theidentification and discussion of cultural norms is something of a commonplacein cross-cultural communication, they are, at best, a shorthand for what isacceptable to an important section of the community, at worst, a crudegeneralization or stereotype of that community. Speakers in any communitycommand a repertoire of speech styles, and the selections they make from thesearise not only from particular strategic considerations or from a desire to behavein normative ways, but are also constrained by the nature of speakers'experience of what is normative, and how they wish to project themselves in anycontext. That is, speaker' uses of politeness phenomena also say somethingabout them as an individual since individuals vary in the experience they have ofthe norms of a community and also in the extent to which they want to projectthemselves normatively or creatively with respect to any norms of which theyare aware. Thus the diversity found among individuals can relate to a number ofbackground and personality variables, as well as to factors of specific context,and is evident in both native speakers and normative speakers of a language. Inthe latter case, of course, the issue is complicated by different levels ofawareness or uptake of pragmatic aspects of language use (e.g., Norton, 2000;Schmidt, 2001).

What implications are there, then, for advanced language learners seeking toconstruct their social identities in a new language and culture? Although"optimal convergence" (Giles, Coupland, & Coupland, 1991) has been argued asan appropriate goal for interlanguage learners, we can assume neither that thisshould be identical to native speaker patterns of use, nor that these patterns areor can be uniform. Indeed, the data examined here demonstrate that they are not.Moreover, from an interlocutor's perspective, it may be that there is a certainthreshold of tolerance for certain amounts of convergence, but that beyond acertain point, a speaker's attempts to converge too closely on a native speakerperformance may be perceived as patronizing (Giles, Coupland, & Coupland,1991, p. 79). In the words of my teenage son, who was a student at one of the


schools where data was collected, a speaker who passes beyond this thresholdmay be dismissed as a "try hard." Because they are NOT native speakers,Chinese background trainee teachers may not be accepted by their students ifthey construct teaching identities that are too obviously "native speakerlike."For example, they may have difficulty addressing a student as "mate" oremploying frequent colloquialisms to reduce social distance. The challenge forthem is to project an identity that is both acceptable to their interlocutors and tothemselves as individuals.

Viewed from an applied perspective, the findings of this study suggest thatany instruction in the use of mitigation, or any other aspect of language use,needs to tackle a range of issues. First, we need to ensure that instructionprograms take into account and make available for examination sociopragmaticissues of what is valued in communication and communicative style, how rolesare perceived and played out in different communities, what interlocutorexpectations of particular role-relationships might be, and so on in any context.In this way learners can come to understand more about expectations bothwithin their new community and in the communities into which they weresocialized as youngsters. Second, there needs to be a focus on the pragma-lingistic devices available for use and their force in context, so that learners canincrease their awareness of the different devices available and how theyfunction. Third, as we have seen, individuals make differential use of the reper-toires of devices at their disposal, and exploration of this individual variationcan be useful, not only in demonstrating the range of behaviours found in nativespeaker performance, but also in illustrating the creative ways in whichindividuals can draw upon their linguistic resources to project a social identitythat accurately reflects their intentions.

REFERENCES

Amin, N. (1997). Race and the identity of the nonnative ESL teacher. TESOL Quarterly, 31, 580-582.


Blum-Kulka, S. (1982). Learning how to say what you mean in a second language: A study ofspeech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-59.

Blum-Kulka, S. (1985). Modifiers as indicating devices: The case of requests. TheoreticalLinguistics, 12, 2-3.

Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal ofPragmatics, 11, 131-146.

Blum-Kulka, S. (1989). Playing it safe: The role of conventionality in indirectness. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies.Norwood, NJ: Ablex.

Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests andapologies (Vol. XXXI). Norwood, NJ: Ablex.

Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge,England: Cambridge University Press.

92 Yates

Busher, S. (1997). Language and cultural identity: A study of Hmong students at the postsecondarylevel. TESOL Quarterly, 31, 593-602.

Cortazzi, M., & Jin, L. (1996). Cultures of learning: Language classrooms in China. In H. Coleman(Ed.), Society and the language classroom. Cambridge, England: Cambridge University Press.

Deby, J. (2003). Masculinity and the place of physical aggression in the structure of Canadiantelevised ice-hockey commentary. Paper presented at the Eighth International PragmaticsConference, Toronto, Canada.

Eslamirasekh, Z. (1993). A cross-cultural comparison of the requestive speech act realisationpatterns in Persian and American English. Pragmatics and Language Learning, 4, 85-103.

Fen, H. L. (1996). Requests and apologies: A study in cross-cultural pragmatics of native Chinesespeakers, ESL Chinese speakers and native Australian English speakers. La Trobe University,Melbourne, Australia.

Ferguson, C. (1975). Towards a characterisation of English foreigner talk. AnthropologicalLinguistics, 17, 1-14.

Fukushima, S. (1990). Offers and requests: Performance by Japanese learners of English. WorldEnglishes, 9,317-325.

Giles, H., Coupland, J., & Coupland, N. (Eds.). (1991). Contexts of accommodation. Cambridge,England: Cambridge University Press.

Goffman, E. (1956). The nature of deference and demeanour. American Anthropologist, 58, 473-502.

Gumperz, J. J. (1982). Discourse strategies. Cambridge, England: Cambridge University Press.Gumperz, J. J. (1996). The linguistic and cultural relativity of conversational inference. In J. J.

Gumperz & S. C. Levinson (Eds.), Rethinking linguistic relativity. Cambridge, England:Cambridge University Press.

Ha, C. T. (1998). Requests by Australian native speakers of English and Vietnamese learners ofEnglish: A cross-cultural communication study in pragmalinguistics. La Trobe University,Melbourne, Australia.

Hansen, J. G., & Liu, J. L. (1997). Social identity and language: Theoretical and methodologicalissues. TESOL Quarterly, 31, 567-576.

Hartford, B., & Bardovi-Harlig, K. (1992). Experimental and observational data in the study ofinterlanguage pragmatics. In L. Bouton & Y. Kachru (Eds.), Pragmatics and languagelearning: Monograph series Volume 4, University of Illinois at Urbana-Champaign.

Hinkel, E. (1996). When in Rome: Evaluations of L2 pragmalinguistic behaviors. Journal ofPragmatics, 26, 51-70.

Hofstede, G. (1986). Cultural differences in teaching and learning. International Journal ofInter cultural Relations, 10, 301-320.

Hofstede, G. (1994). Cultures and organisations. London: Harper Collins.Holmes, J. (1995). Women, men and politeness. Harlow: Longman.Huang, M.-C. (1996). Achieving cross-cultural equivalence in a study of American and Taiwanese

requests. Unpublished doctoral thesis, University of Illinois, Urbana-Champaign.Kim, J. (1995). "Could You Calm Down More?": Requests and Korean ESL learners. Working

Papers in Educational Linguistics, 11, 67-82.Koike, D.-A. (1994). Negation in Spanish and English suggestions and requests: Mitigating effects?

Journal of Pragmatics, 21, 513-526.Kubota, M. (1996). Acquaintance or fiancee: Pragmatic differences in requests between Japanese

and Americans. Working Papers in Educational Linguistics, 12, 23-38.Kuiper, K. (1991). Sporting formulae in New Zealand English: Two models of male solidarity. In J.

Cheshire (Ed.), English around the world: sociolinguistic perspectives. Cambridge, England:Cambridge University Press.

Lee-Wong, S. M. (1993). Requesting in Putonghua: Politeness, culture and forms. Unpublisheddoctoral thesis, Monash University, Melbourne.

Lee-Wong, S. M. (1994). Qing/please - a polite or requestive marker?: Observations from Chinese.Multilingua, 13, 343-360.

Mir, M. (1995). The perception of social context in request performance. In L. F. Bouton (Ed.),Pragmatics and language learning, Monograph series Volume 6, University of Illinois atUrbana-Champaign.


Norton, B. (2000). Identity and language learning: Gender, ethnicity and educational change.Harlow, England: Longman.

Ochs, E. (1988). Culture and language development: Language acquisition and languagesocialization in a Samoan village. Cambridge, England: Cambridge University Press.

Ochs, E. (1993). Constructing social identity: A language socialization perspective. Research onLanguage and Social Interaction, 26, 287-306.

Ochs, E. (1996). Linguistic resources for socializing humanity. In J. J. Gumperz & S. Levinson(Eds.), Rethinking linguistic relativity. Cambridge, England: Cambridge University Press.

O'Driscoll, J. (1996). About face: A defence and elaboration of universal dualism. Journal ofPragmatics, 25, 1-32.

Rintell, E. M., & Mitchell, C. J. (1989). Studying requests and apologies: An inquiry into method. InS. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests andapologies (Vol. XXXI). Norwood, NJ.: Ablex.

Rose, K. R. (1990). Cross-cultural pragmatics: Requests and apologies. Issues and Developments inEnglish and Applied Linguistics, 5, 107-115.

Rose, K. R. (1992). Speech acts and questionnaires: The effect of hearer response. Journal ofPragmatics, 17,49-62.

Rose, K. R. (1994). On the validity of the discourse completion tests in nonwestern contexts. AppliedLinguistics, 15, 1-14.

Scarcella, R., & Brunak, J. (1981). On speaking politely in a second language. International Journalof the Sociology of Language, 27, 59-75.

Schmidt, R. (2001). Attention. In P. Robinson (Ed.), Cognition and second language instruction.New York: Cambridge University Press.

Scollon, S. (1999). Not to waste words or students: Confucian and Socratic discourse in the tertiaryclassroom. In E. Hinkel (Ed.), Culture in second language teaching and learning. Cambridge,England: Cambridge University Press.

Searle, J. R. (1975). Indirect speech acts. In P. Cole & J. L. Morgan (Eds.), Syntax and Semantics 3:Speech Acts (pp. 59-82). New York: Academic Press.

Shea, D. (1994). Perspective and production: Structuring conversational participation across culturalborders. International Pragmatics Association, 4, 357-389.

Siegal, M. (1996). The role of learner subjectivity in second language sociolinguistic competency:Western women learning Japanese. Applied Linguistics, 17, 356-382.

Tang, C. (1997). The identity of the normative ESL teacher: On the power and status of normativeESL teachers. TESOL Quarterly, 31, 577-579.

Trosborg, A. (1995). Interlanguage pragmatics: Requests, complaints and apologies. Berlin:Mouton de Gruyter.

Weedon, C. (1987). Feminist practice andpoststructuralist theory. Oxford, England: Blackwell.Yates, L. (2000). "Ciao, guys!": Mitigation addressing positive and negative face concerns in the

directives of native-speaker and Chinese background speakers of Australian English.Unpublished doctoral thesis, La Trobe University, Melbourne.

Yu, M.-C. (1999). Universalistic and culture-specific perspectives on variation in the acquisition ofpragmatic competence in a second language. Pragmatics, 9, 281-312.

Zhang, Y. (1995). Strategies in Chinese requesting. In G. Kasper (Ed.), Pragmatics of Chinese asnative and target language. Hawai'i: University of Hawai'i Press.

94 Yates

APPENDIX ASummary of Strategy Coding Framework

STRATEGY

Substrategy Socialdistance

Examples

1. Nonexplicit negotiable strategy

State non-conventional

Question non-conventional

State conventional

Question conventional

Mixed

No?

Mixed

Mixed

Now everybody is waiting for you. (=hurry up)(Cf2)

OK, 1 who's our reporter in the back group? 1John? (Aft)

so now's the chance to get it down. (Afl)

OK,/ are you copying this down? (Cml)

2. Apparently negotiable strategy

Permission

Suggestory formulae

Question ability

Willingness question

Yes

Depends

Yes

NoYes

alright lcan I have your attention this way(AmS)

why don't you just write down somewhere/ (Cml)

(name) lean you please get it out? (Af3)

a) so do you want to start, (Cfl)b) (name)/ would you please sit down? (Cm2)

3. Apparently advisory strategy

Willing state

Ability state

Suggest state

Advise state

No

No

Mixed

Mixed

you want to write it down too Natalie (Am1)

you can copy this down into your books .. (Afl)

I think it's best/ now, /we just get stuck into theprac (Am1)

all eyes should be on books (Am4)

4. Apparently assertive strategy

Wants/Needs

Obligation

Predictive

Teacher assert

Elliptical

Imperative

You assert(+ stem or will)

No

No

NoDepends

Depends

Mixed

Yesyes

Now/ I want the boys all to move /thata way/okay (Cm3)

You '11 have to move that car (Blum-Kulka,Kasper & House, 1989, p. 279)

alright /so first we're going to listen (Aft)

now this is homework,/ yeah (Cm2)

back crawl /not breast stroke (Am5)

Clean up that mess (Blum-Kulka, Kasper &House (1989)

a) you go on with your work (Holmes, 1983 :99)b) you are here. 1 you will start from here (m)(Cm21:l)


APPENDIX BSummary of Mitigation Coding Framework

SYNTACTIC MITIGATION

Syntactic device

Past tense

Unrealisers Conditionalsfull/suppressed

Modals

Future

going to/will

Deflectors Defocaliserspseudo-clefting

Continuous

Social distance

Mostly

Depends on use

Depends

No

No

No

Example

Could you hand me the paper please?(Trosborg, 1995, p. 210)

now, if you can get into groups oftwo, (Af4)

all right /may I have your attentionplease (Cf5)

you are going to have to stopdistracting all the people around you(Af3)a) so, what I suggest you do/ is justread through it, (4f3)

treading water guys,/ not floating,(AmS)

PROPOSITIONAL MITIGATION

1. Nonimpositive prepositional mitigation

Example

Reason

Preparator

Conditions

Sequence

Examples

Underplay

I missed class yesterday. Could I borrow your notes? (Blum-Kulka etal., 1989)

OK, / Got everything? FS: No T: Just hurry up. / Chop chop(Cml)

now what I'd like you to do is /just heat that up /we 've gotthermometers here, (Am1)

try it again with the fifty /and /'// see if I can find anothercalorimeter (Am1)

and you work out what's the answer to this one (for example,/ what'sthat fraction? (Af3)

i) Look I'm sorry/, I don't wanna have to keep going on/ at your back/but /but otherwise I'm going to have to move you/ OK? (m)(Af11:12)ii) and (name)/ you can report if you like, (=be reporter in yourgroup) (m) (Af21:7)

96 Yates

APPENDIX B (continued)

2. Positive propositional mitigation

Encouragement

Personal or culturalreference

Humour

1. Please

Requestive politenessmarker, Please

Ok that was good /Come in here guys (Am5)OK,/well,/ just go on with chapter nine/ don 't worry about it. OK.(Af3)get off it (name) /it's not a surfboard alright? (Am5)

how have we gone three B./ Finished? T: You lucky girl. 1 T: comeon. (Af1)

LEXICAL MITIGATION DEVICES

Hand me the paper, please (Trosborg, 1995, p. 212)

2. Minimising negative devices

(i) Downtoner (just)(ii)UnderstaterEpistemic uncertainty/undermining hedge

3. Warmth and apprc

Agreement

Token tag

i) and ii)just oh sorry () can everyone just () listen for a minute.(Am1)

Let's maybe look at this one / think (Aft)

ival positive devices

i) yes, /write this down. (Cm4)

Clean up the kitchen, okay! (Blum-Kulka, Kasper, & House, 1989,p. 285)so now /you/ you look at it this yourself /you understand! (Cf1)


APPENDIX CUse of the Four Categories of Mitigation by Individual Participants

Participant Indirectstrategies/total

directives

AflAf2Cf1Cm1Am1Af3Cf2

Am2Af4Cf3

Am3Am4Am5Cm2Cf4Cf5Cm3Cm4

.510

.346

.438

.361

.189

.279

.241

.320

.262

.282

.136

.274

.084

.113

.119

.092

.091

.028

Syntacticmitigation/total

directives

.204

.167

.087

.123

.114

.130

.065

.056

.171

.051

.121

.104

.031

.017

.005

.015

.044

.009

Lexicalmitigation/total

directives

.520

.468

.427

.445

.538

.498

.540

.320

.276

.322

.424

.301

.454

.379

.356

.262

.228

.343

Propositionalmitigation/total

directives

.459

.276

.272

.258

.326

.229

.177

.304

.211

.203

.167

.120

.200

.119

.124

.165

.097

.037

Mitigationindex

1.691.261.22

1.19

1.17

1.141.02

1.00

.92

.86

.85

.80

.77

.63

.60

.53

.46

.42

4

Before, During, and After the Event:Getting the Job (or Not) in an Employment

Interview

Julie KerekesCalifornia State University, Los Angeles

Advances in sociolinguistic investigations of intercultural gatekeepingencounters have evolved from those which attribute miscommunication and theresulting failure of such encounters to cultural mismatches (e.g., Akinnaso &Ajirotutu, 1982; Gumperz, 1982, 1992), to a more complex recognition thatthere is indeed no one-to-one correspondence between cultures andcommunicative styles. More recently, scholars have suggested that shareddiscourse styles (Gee, 1996) and an establishment of rapport through co-membership (Erickson & Shultz, 1982) contribute to a matching of communi-cative styles. It is these matches, they claim, rather than the more simplisticview, which, similar to previous claims about cultural matches, can lead tosmooth verbal interactions. Moreover, further consideration has led scholarssuch as Shea (1994) and Meeuwis and Sarangi (1994) to recognize the role thatpretext—the greater context of an interaction, including structural parameters,power dynamics, dominance, and prejudice—plays in influencing the potentialfor a successful gatekeeping encounter.l A number of similar studies have alsoindicated that one main determiner in the outcome of a gatekeeping encounter isthe establishment (or lack thereof) of a trusting relationship between the inter-locutors involved in that gatekeeping encounter (Gumperz, 1992; Kerekes,2003); that is, the relationship between the person being judged and the judgingauthority figure whose legitimate, institutionalized role enables her or him tomake a decision which will effect the future of the person she or he judges.Although scholarly perspectives on gatekeeping encounters have thus widenedto include factors not evident merely in the verbal interactions themselves, thefocus of most sociolinguistic studies on this topic still lies primarily on theverbal actions which occur during the gatekeeping encounter, and/or nonverbalfactors immediately and directly affecting those verbal actions. It is the aim ofthis chapter to consider, therefore, to what degree factors outside of thegatekeeping encounter itself—even those that do not overtly or directly comeinto play during the gatekeeping encounter—affect its outcome.

1 For the purposes of this discussion, I have adopted Schiffrin's (1994) definition of gatekeepingencounters: "asymmetric speech situations during which a person who represents a social institu-tionseeks to gain information about the lives, beliefs, and practices of people outside of that institu-tionin order to warrant the granting of an institutional privilege" (p. 147).

99

100 Kerekes

The data source for this chapter is a set of job interview transcripts andrelated data collected at a San Francisco Bay Area branch of an employmentagency which I will call FastEmp.2 Forty-eight job candidates and four staffingsupervisors (who interviewed them) participated in this study, by allowing theirinterviews to be observed and audio- and video-recorded, and by partaking infollow-up and debriefing interviews with the researcher after they hadcompleted their job interviews.

THE INSTITUTION

Although a candidate's qualifications are purportedly evaluated by means of thejob interview and the documents accompanying the interview (resumes, jobapplications, test scores), in reality the staffing supervisor's assessment of thejob candidate derives from a complex web of information obtained outside andinside the interview, including before, during and after the interview. Therefore,on a more macro level, in this chapter assessments are made on the staffingsupervisor's—and her institution's—stereotypes of expected behavior on theparts of particular subgroups of job candidates. It is necessary, therefore, toconsider the greater context in which these interviews occur.

Employment Agencies and the Job Market

Employment agencies play an indispensable role in the American economy andwork culture not only because of the high revenues they bring in—in 1998alone, they brought in $72 billion—but also because of the unique purposes theyserve.3 When an employer has a sudden need to increase the workforce, doesnot have the time or resources to employ a new worker directly, or otherwisedoes not want to take on the responsibilities of permanently hiring a newemployee, the employer seeks assistance through such employment agencies(i.e., the staffing industry). When an individual seeks immediate employment,has commitments which leave her or him unable to go through the (oftenlengthy) process of applying for permanent work, or is otherwise unable toobtain a permanent job, she or he can apply for a job through an employmentagency and, often, begin working the same day. In general, employmentagencies serve as the inter-mediate agent in a transaction between a clientcompany and a potential worker (a job candidate).

Client companies use staffing services in situations of temporary skillshortages, seasonal workloads, employee absences, and special assignments orprojects. They place their order with an employment agency, specifying thequalifications they seek in a job candidate; it is then the job of the employmentagency to find a candidate who meets these requirements. At the same time,

2 This chapter draws from a larger study discussed in Kerekes, (2001, 2003a, 2003b).3 In addition to the sources cited in this section, employment agency statistics discussed in thischapter were obtained from sources listed in footnote 6.

Getting the Job 101

individuals regularly seek jobs through employment agencies by presenting theirskill sets, stating their job requirements (in an interview), and applying for amatch through the employment agency.

Not only does the staffing industry have an impact on the efficacy andsuccess of a variety of businesses and other organizations, it also touches animpressively diverse range of the U.S. population, serving every socioeconomicclass. Companies ranking highest in the Fortune 500 all the way down to one-person businesses, as well as organizations in the nonprofit sector, rely onemployment agencies to provide temporary work services. Whereas manytemporary positions require highly educated and technically skilled candidates,others require skills not learned in the classroom, and often performed byworkers with very little formal education or training. Temporary workersrepresent the whole spectrum of educational backgrounds, from those with lessthan a high school education to others with multiple advanced degrees. Theyrepresent a range from young, novice workers seeking work experience toupgrade their resumes, to veteran workers with extensive and variedprofessional experience. They represent the United States' myriad ethnicities,classes, and sociolinguistic backgrounds.

Given their crucial role in the U.S. employment industry, accuratelyassessing their job candidates and making correspondingly successful jobplacements are, therefore, crucial matters taken seriously by employmentagencies. Employment agency staff undergo extensive training—the mostimportant parts of which address the job interviews they conduct—to improvetheir hiring practices, with the aim of making successful matches between clientcompanies and job candidates.

FastEmp

FastEmp is a national employment agency with approximately 200 branchesacross the United States. Silicon Valley's South Bay branch of FastEmp(henceforth referred to as "FastEmp"), the site of this study, serves jobcandidates who represent diverse career interests and a wide range of languageand educational backgrounds. The majority of normative speaker (NNS) jobcandidates undergo FastEmp placement procedures in English, but some areassisted by bilingual interpreters (these are usually family members or friendswho accompany the candidates to their appointments at FastEmp).

The FastEmp staff are aware of the preponderance of culturally andlinguistically diverse job candidates at this branch, as well as of FastEmp'sreliance on the NNS segment of the population to fill many of their positions,especially those in the light industrial sector. For this reason, they are generallyenthusiastic about doing what they can to improve the compatibility of theirhiring procedures with the needs of their NNS job applicants. Regionalsupervisors of FastEmp have made a call for innovative developments to bettermeet the needs of their linguistically diverse clientele. In contrast to theirclientele, the South Bay FastEmp staff is much more socioeconomically

102 Kerekes

homogeneous: European-American women born in the United States, nativespeakers of English, and self-described as middle class, with two or more yearsof higher education.

THE JOB INTERVIEW

It is through the job interview that the staffing supervisor assesses thecandidate's qualifications for work opportunities with FastEmp. Qualificationsinclude but are by no means limited to the actual job skills a candidate possessesand is able to exhibit. In addition, and of great importance to the staffingsupervisor, is her evaluation of how well the candidate communicates, and whatkind of impression she feels this candidate would make on potential clients invarious job settings. She assesses these qualities on the basis of her perceptionof the job candidate's flexibility, ability to learn quickly, and trustworthiness.Finally, from the staffing supervisor's perspective, the job interview is heropportunity to make the possibility of employment at FastEmp attractive to thecandidate, especially if she has a positive impression of the candidate. For thejob candidate, the interview is also potentially a time to determine what the workoptions are and to assess whether they are worth taking.

Components of the Job Interview

The interview is generally comprised of a fairly routinized question and answersequence, structured to include the following components: introductions, workpreferences, work qualifications, wrap-up. In addition to basing her assessmentof the job candidate on the verbal answers supplied by the candidate, the staffingsupervisor refers to the candidate's completed application form and/or resume,test scores, and conversations she may have had with the candidate'sprofessional references.

Introductions. After a brief greeting, the staffing supervisor goes through ashort series of rote questions, verifying the job candidate's address, phonenumber, and other information the candidate has supplied on the writtenapplication. She asks the job candidate how she or he found out aboutFastEmp's service and whether her or his attention was drawn to a particularadvertisement or current job description. While on the surface purely practical—FastEmp must have accurate contact information for the job candidates in orderto eventually place them on assignments and pay them—these questions alsogive the staffing supervisor her first impression of the candidate's social andliteracy skills (What does her/his written application look like? Is her/hiswriting legible? Are there spelling errors? Have the printed questions beencorrectly interpreted?). The following excerpt from an interview between Erin(staffing supervisor) and Frank, a Spanish L1 light industrial candidate,exemplifies the introduction segment (lines 1—28). Frank is interested in a

Getting the Job 103

specific job order that has been placed with FastEmp, to do temporarywarehouse work at a company called Eagle:4

(1) Erin and Frank (L1 Spanish-speaker, light industrial candidate)

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354

Erin

FrankErin

FrankErin

FrankErinFrankErin

FrankErinFrankErinFrankErin

FrankErin

FrankErin

FrankErin

FrankErinFrankErin

Frank

ErinFrank

ErinFrankErin

FrankErin

Now you were referred by. . . who is it that you knewat xx in um, at Ea[gle?]

[Uh, ] Ernesto.Ernesto, okay (3.0) great, um, and you've alreadyspoken with Oscar, is that correct?Yeah.Okay (3.0) terrific. What we're gonna do is we'regonna confirm the information on your application,and um talk a little bit about what it is um ideallythat you're looking for 'n specifically about Eaglesince you're here for==0h.==the Eagle position.Yes.Um, talk a little bit about the references that Ican call, and then we'll go from there.Oh.Um . . . you're at three twenty-one third lane?Yeah.Is that correct in South San Francisco.MhmAnd at this home phone number, six five four fourone two eight, is there an answering machine at thatnumber?Yes.Okay. (2.0) And this other number, what kind of..[contact is][Uh, that's] .. my dad's house.Okay. Great. Um. And what are you looking forhourly?Hourly, it doesn't really matter, whatever, e[ight.]

[Okay.]Eight plus okay. Um Eagle typically starts at nineplus, ==oh==so I'll just put that on there, if you don't mind.Yeah.Um, and you're looking for part-time?[Are] you look [ing] for evenings or [or full time?][uh] [or] [full time]yeah.Okay. What hours are you looking to work, ideally?Um, just basically prob'ly from .. seven to when Ican I mean night, midnight.Seven p.m. ?

Yes . [xx][Okay] ...to midnight-ish. Okay. Um, an:d

{ ( (reading his application) ) looking for fast pacedenvironment}, you'll get that at Eagle. Um, yourhigh school diploma, you have not

re[ceived it?]{((shakes head)) [No.]}

{( (writing) ) Okay} . (2.0) Mkay. So what brings you toEagle?

4 This and all other company and employee/employer names are pseudonyms.

104 Kerekes

Work Preferences. In the work preferences component of the interview,the staffing supervisor ascertains details from the candidates about time,location, and other constraints the candidates may have on types of work theyare willing to do. She asks the job candidates how far they are willing to travel,their expected salary, whether they seek temporaiy or permanent work, andwhen they are available to begin work. This series of short-answer questionsprovides the candidates with opportunities to show the staffing supervisor howflexible or inflexible they are regarding the range of assignments they arewilling to take. During this time it also becomes apparent to the staffingsupervisor how familiar the candidate is with the institutional job placementprocedure: Do they have ready answers to her routine questions, exhibiting aknowledge of how employment agencies function? Do they understand whatsorts of answers are expected of them? Finally, how clearly can they answer thestaffing supervisor's more open-ended question, "What is it ideally that you'relooking for?" An example of the work preferences segment of the job interviewcan be found in Frank and Erin's interview (Example 1), lines 29-54. In thissegment Erin asks Frank a series of questions including "What are you lookingfor hourly?" (line 29), "You're looking for part-time?" (line 38), "What hoursare you looking to work, ideally?" (line 42). She also helps him fit his answerabout hours more closely to what is expected at the company he is interested in(lines 32-37).

Work Qualifications. Discussion of what type of work the job candidateseeks generally segues into the third component of the interview, in which thecandidate's work qualifications are discussed. In this section, the staffingsupervisor verifies the job candidate's employment history, related experience,and educational background. The framework for this discussion is achronological account of the candidate's work history, using jobs listed on theapplication and/or resume as cues. There is great variation as to which detailsare attended to, however, depending on the staffing supervisor's impression ofthe candidate thus far. While this component may include detailed description ofactual work done and skills learned, it may also focus solely on precise datesand locations of previous assignments. It is in this part of the interview that acertain level of trust or distrust manifests itself between the interlocutors. Theextent to which the staffing supervisor questions the job candidate about exactdates and locations of previous jobs hinges on whether she finds the jobcandidate credible on the basis of what she has seen up to this point. The themeof trustworthiness is associated with the potential which this candidate may haveto complete a job assignment successfully (see Kerekes, 2001, 2003a, 2003b). InExample 2, we see part of the work qualifications segment of a job interviewbetween Carol, a staffing supervisor, and Lisa, a native-speaking AfricanAmerican candidate for light industrial work.

Getting the Job 105

(2) Carol and Lisa (NS, light industrial candidate)

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455

Carol

LisaCarolLisaCarolLisaCarol

Lisa

CarolLisaCarol

Lisa

Carol

Lisa

Carol

LisaCarol

LisaCarol

LisaCarol

Lisa

CarolLisaCarolLisa

An:d you're currently working at NorinProduct [s?]

[Mhm.] [Yes.][Doing assembly] work?

Yes .Okay um:, and you've been there since February?Mhm.Kay. Building a variety of cabinet coolers forcustomers .Cabinet coolers um they cool electrical systems in.. in machinery., from big heavy duty fans, to, it'sjust a a coolant system.Do you enjoy that kind of work?No [ ( (laughs) ) ] .

[No. Okay.] Okay. I can understand that I geta lot of people that find um assembly work a littletedious?==The assembly is fine because I'm I'm mainly used todoing medical assembly, technical assembly, theelectronic assembly, (.hhh) and this is just Ithought it was gonna be all right but when theyhired me (-hhh) they hired me to do testing and I'mnot doing testing I'm doing everything I'm cleaningfloors I'm doing everything else but testing. ==0kay. (5) So what is it that you do enjoy doingLisa(•hhh) Um, I I like the assembly aspect because Ilike the speed of it sometimes their speed um, thetechnical aspect of it. You have to be veryprecise, uh very detail oriented (3) veryflexible ((laughs))(3) Great. Is there anything you just really don'twant to do?==Telemarketing ( (laughs) ) .That's good for us to know. We will- we'd docu-document that in our system, ==0h, o [kay] .

[and then] that way we would never call youfor that [type] of position.

[Okay.] Right.What would a former employer tell me about your workperformance?Um that I'm that I'm very reliable, that I show up,{((laughing)) on time}. A::nd uh that even though Ihave a a bike um I always make arrangements ifthere's if I hear that there's we're gonna havereally bad weather I always try to make arrangementsway in advance to let them know that I'm gonna belate if I'm gonna be late.Ever had any attendance issues on any of your jobs?==Never. The only time I'm ever out is if I'm sick.And what are your long term goals?Long term goals I'd love to find a company that Ican stay with and grow with and stay with for . .many years and move up in the company.

106 Kerekes

Wrap-up. The wrap-up component consists primarily of a monologue onthe part of the staffing supervisor, who explains in varying amounts of detailwhat the next steps are for the job candidate (depending on her or his potentialto be placed on an assignment). She may have the candidate fill out legal forms,and walk the candidate through an orientation packet with information aboutcompany policies. In addition, if there is currently an open position the staffingsupervisor feels the candidate may be able to take, she describes it to her orhim.5

Upon completing the interview, the staffing supervisor makes anassessment of the candidate and has a fairly clear idea as to whether or not thecandidate is hirable. In many cases, the staffing supervisor has not yet finishedchecking the job candidate's references, however; these may result in herchanging her assessment of the job candidate.

THE PARTICIPANTS

Job Candidates

The 48 job candidates in this study are equally distributed across gender (24females and 24 males) and job type (24 light industrial and 24 clericalcandidates). The job candidates are also equally distributed according to theirnative language status (24 NSs and 24 NNSs). NSs self-identified as such, wereraised speaking English, and named English as the language in which they feltmost proficient. NNSs were raised speaking a language other than English andself-identified as normative speakers.6 Four of the 48 job candidates self-identifyas bilingual. Of these, two consider their native language to be English and arewidely recognized as native speakers of English; the other two speak Englishwith phonological features of Spanish and are seen by their interlocutors (i.e.,the staffing supervisors in this study) as NNSs.

All of the NNSs in this study, as well as 11 of the NSs, are people of color,self-identified as well as determined on the basis of appearance. In terms of thefive races recognized by the U.S. Census Bureau,7 with the addition of "Latino"as a race, 13 of the NNS participants are Asians, 9 of the NNSs are Latino, and 2of the NNSs are Pacific Islanders. In total, of the 48 participants (including

5 The staffing supervisors requested that this segment of the job interview not be recorded for thisresearch project, in order to protect the privacy of the candidates, who often filled out legal formsduring this concluding part of the interview.6 In total, participants represent eleven native languages: Cantonese, English, Hindi, Japanese,Mandarin, Samoan, Spanish, Tagalog, Tongan, Urdu, and Vietnamese.7 The U.S. Census Bureau designated the following as race categories in the 2000 Census: White;Black or African American; American Indian and Alaska Native; Asian; Native Hawaiian and OtherPacific Islander; and "Some other race." The question of Hispanic (Spanish/Hispanic/Latino) originis asked separately in the 2000 Census.

Getting the Job 107

Table 4.1. Job Candidates by Gender, Job Type, Native Language, and Race

Job type Gender

Female

Light industrial(total 24)

Clerical(total 24)

Language status

NNSNS

NNSNS

Background

5 (2A,2L,1P)5 (3B, 2W)

5 (3A, 2L)9 (3B, 1L, 5W)

Male

Language status

NNSNS

NNSNS

Background

9 (4A, 5L)5 (3B, 1L, 1W)

5 (4A, 1P)5(5W)

Total 24 Females 24 Males

Note. A = Asian; B = Black/African American; L = Latino; P = Pacific Islander; W = White/Euro-pean American

NSs and NNSs), 13 (27%) are Asians, 9 (19%) are Black, 11 (23%) are Latino,2 (4%) are Pacific Islanders, and 13 (27%) are white.

The racial distribution across job type for male participants is not equal, butrather a realistic representation of the general distribution of such jobs amongFastEmp employees: All of the NS male clerical candidates are white, while 4out of 5 of the NS male light industrial candidates are Black or Latino. Amongthe female candidates, the racial distribution represents a closer balance acrossjob type, again realistic for FastEmp's general patterns of job distribution. Thesenumbers are summarized in Table 4.1.

The job candidates' ages range from 18 to mid-60s, with more than twothirds of them in their 20s to 40s. Educational backgrounds (highest gradecompleted) range from not finishing high school to completing a master'sdegree or medical degree, or currently working on a Ph.D. The majority of thejob candidates have completed some post-high school course work but have notobtained a higher degree. Many of these candidates have taken vocationalclasses (e.g., electronics, welding). Figure 4.1 shows the distribution of educa-tional backgrounds of the job candidates, with a general trend toward the lefthalf of the graph for light industrial candidates, that is, candidates whose highestcompleted level of education is high school or just a few classes (but no degree)beyond high school, and toward the right half of the graph for clericalcandidates. All clerical candidates have graduated from high school, and moreof them have taken college classes or received higher degrees than of their lightindustrial counterparts.

Staffing Supervisors

Three full-time staffing supervisors, Amy, Carol, and Erin—NS White, self-proclaimed middle class women between the ages of 25 and 30—were

108 Kerekes

Figure 4.1. Job Candidates' Educational Backgrounds

employed at FastEmp during the data collection phase of this study (but not allconcurrently, due to a high turnover rate). The fourth, Renata, was "borrowed"from a different Bay Area branch of FastEmp on a few occasions, to help out onthe days they were understaffed. Also female and in her late 20s, Renata is ofLatin American descent and a native speaker of English; she conducted one jobinterview which is used in this study.

DATA SOURCES

Over a 14-month period, a combination of seven data sources was used toinvestigate when and why a job candidate succeeded or failed to be eligible forwork with FastEmp; each source contributed different kinds of information, andalso, together, provided a means of triangulation (Denzin & Lincoln, 1994).They included the following sources, each described below:

1. Introductory interviews with FastEmp staff.2. Job interviews for prospective employees.3. Follow-up interviews with job candidates.4. Debriefing interviews with FastEmp staffing supervisors.

Getting the Job 109

5. Final interviews with staffing supervisors.6. Site documents.7. Field notes.

Introductory interviews with FastEmp staff were conducted at the beginningof the study, in order to get acquainted with them and their institution. Thequestions I asked served to determine the typicality of the South Bay branch ofFastEmp, in relation to other FastEmp offices, as well as in relation to otheremployment agencies. My questions focused on: (a) the procedures routinelycarried out at FastEmp; (b) the job placement process as it affects both thestaffing supervisors and the job candidates; (c) the degree to which staffingsupervisors are aware of issues involving intercultural communication with jobcandidates and/or clients of diverse cultural/linguistic backgrounds; and (d)identification of characteristics of prototypical job candidates and clients.Further questions addressed the staffing supervisors' philosophies andperspectives on the job interviewing process; how they find quality candidatesand what constitutes a quality candidate; and their attitudes about candidates'skills, their language use, and their L2 abilities (where relevant).

Forty-seven job interviews were observed, recorded (audio- and video-taped), and transcribed. They were analyzed, using an interactionalsociolinguist-tic approach, to illustrate how successful and failed job interviewswere coconstructed.

Thirty-seven follow-up interviews with job candidates were conducted afterthe candidates had completed their job interviews and all other FastEmp-relatedbusiness (i.e., filling out papers and receiving instructions). These semi-structured interviews were designed to determine the job candidates' perceptionsof the interactions that had taken place during the job interview; to comparetheir intrepretations to those of the staffing supervisors; to learn about theirexpectations of job placement possibilities and attitudes toward looking for workand to learn about their linguistic and educational backgrounds as well as theirknowledge of the employment agency interviewing protocol. Of particularimportance were questions focusing on the candidates' awareness of what theythought their strengths were, how they tried to "sell" themselves in theinterview, and what qualities they thought the interviewer was looking for.

Forty-two debriefing interviews with staffing supervisors were heldimmediately after they finished interviewing the job candidates. These weredesigned to discern their overall impressions, including their assessment of thejob candidates' skills, qualifications, language use, and interviewing competence(as manifested in this particular job interview).

Final interviews with staffing supervisors were conducted at the conclusionof the study. As the data collection progressed, the FastEmp staff becameincreasingly familiar with my research and the sorts of questions I was interestedin investigating. On occasion, they volunteered tidbits they thought would helpme, such as intercultural encounters they had had, their experiences being video-taped in job interviews, and information about individual clients and job

110 Kerekes

candidates. My final interviews with them were therefore designed to allowthem to elaborate on their thoughts and stories. They talked about their currentpositions, future career plans, and FastEmp's role in their professionalaspirations. I repeated or rephrased some of the questions I had originally askedin the introductory interviews, and also obtained updated information regardingthe current status of each job candidate who participated in the study. Thestaffing supervisors revealed that some job candidates who had had excellent jobinterviews had since been disqualified for future FastEmp assignments forvarious reasons ranging from perceived attitude problems to failing to performrequired assignments satisfactorily.

Various site documents were gathered containing information about the jobcandidates' work and educational history as well as other socioeconomicinformation. Sources included the candidates' job applications, resumes,placement test scores, and notes taken by staffing supervisors during the jobinterviews. Other site documents provided information about FastEmp policiesand professional goals and FastEmp's institutional culture. These included, forexample, guidelines for receptionists' telephone speaking behavior, instructionsfor how to screen job candidates over the telephone, motivational memos to thestaff about increasing their clientele and the numbers of job candidates theyinterview per week, and sheets of paper with mottos about how to do goodbusiness, such as a "Whatever It Takes" packet, which contains "favoriteinsights" as reminders of FastEmp's "continued commitment to understandingand exceeding your expectations—whatever it takes."

Finally, I kept detailed field notes from observations of activities andinteractions that occurred during the time that I was present at the site. Beforefocusing on the job interviews and events related specifically to them, myobservations served the purpose of acquainting me with FastEmp's objectivesand modes of operation. I observed the physical setting-the office layout,purpose of each space, ways the rooms were used and by whom. I also observedthe everyday patterns and routines of the FastEmp staff. I sat in the back room—in which the copy machine, printer, job candidate files, and kitchen werelocated—and observed the comings and goings of the FastEmp staff. I observedthe ways they related to each other, studied the FastEmp literature, and noticedhierarchical relationships.

By being present not only for the purpose of collecting data, but also as atemporary employee, volunteering to work as a receptionist at the front desk, Iwas exposed to interactions and events I would have missed, had I simply beenthere as a researcher. I noted significant events and interactions eitherimmediately after they occurred or as soon as I left the site, and used them tocontextualize and triangulate my research findings.

I approached the interviews between the participants and myself as acollaborative project in creating knowledge. It became necessary to recognizemy own role in the creation of meaning during my interviews with both the jobcandidates and the staffing supervisors of FastEmp. That is, what they told meabout their experiences in the job interview, as well as about their perspectives

Getting the Job 111

on other topics we covered, was directly influenced by my presence, and by theinteraction that we coconstructed.8

A DAY IN THE LIFE OF FASTEMP

While the focal point for both the candidates and the staffing supervisors is thejob interview, a number of other interrelated activities take place in the FastEmpoffice and influence the actions and interpretations of the participants in the jobinterview. The context for the job interviews is, therefore, better understoodthrough a description of the general procedures of the job candidates andstaffing supervisors on a typical day at FastEmp.

The Job Candidate's Procedure

A job candidate begins the job application process by calling the FastEmpreceptionist on the telephone, faxing in a resume, or walking in to make anappointment. Before scheduling an appointment, the receptionist screens thecandidate's qualifications, inquiring about the caller's specific skills that mightfulfill some of FastEmp's needs for current or potential position openings. Shealso obtains information about the caller's education, previous work experience,and, most importantly, "people skills"—those characteristics which indicate tothe receptionist that this candidate is a potential "good representative" ofFastEmp. If the candidate meets these preliminary qualifications, thereceptionist schedules an appointment generally within the next 2 days. The jobcandidate is told to expect to spend 2 to 3 hours at FastEmp, filling outapplication forms, taking a battery of tests, and being interviewed by a staffingsupervisor.

Approximately 50% of the job candidates who have appointments actuallyappear at FastEmp at the designated time for their interviews. Before theinterview, they fill out a seven-page application form which includesbiographical data, availability to work (time and location requirements orlimitations), type of work sought (short term, permanent, professional, casual),information about education and work history, professional references, and abattery of tests. Light industrial job candidates are tested in proofreading,reading comprehension, arithmetic, filing, and on-the-job safety policies.Candidates for clerical positions take tests in proofreading, spelling, arithmetic,filing, and typing. Depending on their abilities and work interests, clericalapplicants may also be assessed for data processing and word processing skillsbefore or after their job interview.

8 As stated by Holstein and Gubrium (1995), "interviewers are deeply and unavoidably implicated increating meanings that ostensibly reside within respondents" (p. 3). This was as true of me, when Iinterviewed the participants, as it was of the staffing supervisors, when they interviewed the jobcandidates in the job interviews.

112 Kerekes

The job candidate spends anywhere from 20 to 90 minutes filling out theapplication forms and then waits to be called for an interview with a staffingsupervisor. This wait can take from a few seconds to half an hour or more.When the staffing supervisor is ready, she leads the candidate to a conferenceroom where the two seat themselves at one end of a large table. She thencommences the job interview. They have relative privacy in the conferenceroom, where the door is closed, but the walls are glass and there are occasionalinterruptions from the receptionist or other staffing supervisors who step insideto speak with the interviewer.

At the conclusion of the job interview, if the staffing supervisor does notrequire further tests of the candidate, the job candidate is free to go. Candidatesare invited to make appointments to use software training programs available tothem, should they wish to add skills to their repertoire in order to be eligible formore or better paying jobs.

The Staffing Supervisor's Procedure

Staffing supervisors' responsibilities include networking with clients (thosecompanies that seek temporary or permanent employees through FastEmp),interviewing job candidates, and matching qualified candidates with appropriatejobs. They earn a base salary plus commission, which can be as much as 50% oftheir total income. They describe their positions as a stepping stone for youngprofessionals planning to move up the corporate ladder. As Amy explains, "Youcan't do this job for a long time. ... If you don't move up into managementwithin a few years of doing this job, then you're really not gonna go anywherein this industry. It's a burn-out industry."

While the staffing supervisors at South Bay FastEmp follow oneinterviewing protocol, each of them brings an individual gatekeeping style to theinterviews, as well as to the follow-up with the job candidates. They differ, forexample, in how they deal with documentation of the job interviews, in theirattitudes about the importance of job candidates' references, and in averageduration of their interviews. None of them is completely consistent in theirtreatment of the candidates. (Carol, for example, does not always checkreferences before interviewing a job candidate, despite her claim that this is herregular practice). This is to be expected, given the variability of moods,candidates, schedules, and tasks to be carried out on any given day at FastEmp.We see in the following discussion that some candidates are given more leeway,or benefit of the doubt, than others who have similar weaknesses in theirapplications or interactions; this is relevant to the analysis of the distribution ofsuccessful and failed interviews. Findings regarding the staffing supervisors'differences, and how these differences affect the outcomes of their jobinterviews, are discussed later.

BEFORE THE EVENT

Getting the Job 113

FastEmp staffing supervisors have expressed awareness of the linguisticdiversity of their job candidates, and a concern that intercultural miscom-munications during the job interviews with some of their NNS candidates haveresulted in inappropriate job placements. They feel that the time is ripe fordeveloping strategies for better assessing the skills of their candidates byunderstanding more about intercultural miscommunication, so as to find bettermatches for candidates and clients.

While, overtly, the verbal interactions are the focal point of the jobinterviews, they are affected by a myriad of nonverbal variables. Many of theseare before-the-event factors, such as whether or not the candidate arrives to heror his interview punctually, what FastEmp's current needs are, and thepreconceived notions of the interviewing staffing supervisor. Despite theirawareness, in general, of issues of cross-cultural differences in the workplace,the staffing supervisors are not, for the most part, aware of how their ownpreconceived notions about L2 ability, ethnic/cultural background, social class,and type of work for which the candidates are applying influence theirassessments of the candidates' applications and interviews. Their attitudes aboutdifferent types of job candidates—light industrial versus clerical—result in theirfocusing on different questions with different candidates. The clericalcandidates, in general, have more opportunities to demonstrate their areas ofexpertise through content of more open ended questions addressed in theirinterviews. In contrast, interview questions addressed to many of the lightindustrial candidates focus more on when and where they worked in the past,than on what they actually did in their previous positions. The followingsections present examples of these different interview styles. The outcomes, alsodiscussed below, illustrate an imbalance in success rates according to job typesand, correspondingly, according to the racial distribution of the candidates.

DURING THE EVENT

Stylistic Differences

As mentioned earlier, each staffing supervisor takes an individual approach tothe interviews. They differ in the overall amount of time they spend with eachcandidate, as well as in the areas they choose to emphasize.

Erin, whose light industrial interviews lasted 13 minutes and whose clericalinterviews lasted 21.5 minutes on average, took the most time explaining to thecandidates what the purpose of the interview was, giving them an overview ofthe job interview, and asking detailed questions about their work preferencesand backgrounds. Amy, in contrast, had a more direct and expedient approach,which was manifested not only in the way she conducted the job interviews—which lasted on average 9.5 minutes for light industrial candidates and 12.5minutes for clerical candidates—but also in the other tasks she carried out aspart of her job. Sacrificing perfection for efficiency, Amy utilized the leastamount of time for niceties, making less of an effort than her colleagues to build

114 Kerekes

rapport with her interlocutors. Carol, like Erin, displayed more patience with hercandidates, and a desire to help them feel comfortable in the interview. Sheoften deferred to the expertise of Amy, however, because she was new to the joband felt the candidates sometimes asked questions she could not answer. Her jobinterviews were the shortest in length (both her clerical and light industrialinterviews averaged 7.5 minutes), because, she explained, she was lessknowledgeable than the other staff about FastEmp job opportunities. Renatatook the approach of a teacher. In her interviews, after examining their resumes,Renata went over the candidates' strengths and weaknesses with them and madesuggestions for improvements they could make in future versions of theirresumes, in order to sell themselves as more attractive job candidates.

Verbal Evidence of Success or Failure

The verbal interactions of the job interview generally follow the sequence ofcomponents detailed earlier. Within these sections, a number of strategies and/oractions occur which can "make or break" the encounter. One of the criticalverbal actions that can potentially lead to either a successful or failed jobinterview is the discussion of reasons for leaving previous places ofemployment. Depending on how the candidates choose to explain why they lefttheir previous positions, they either arouse the suspicion of their interviewingstaffing supervisor or assure the supervisor of their responsible conduct. In otherwords, beyond giving the staffing supervisor truthful answers to her questions,the job candidates are variably successful in offering acceptable reasons fortheir answers.

Mark. In the case of Mark, a job candidate for light industrial work,distrust is established when he is asked by Amy (the staffing supervisor whointerviewed him) to explain why and when he left a particular job listed on hisapplication; consequently, Mark fails his interview. Mark is a fluent NNS,having completed his primary and secondary schooling in an English-speakingschool in his home country, the Philippines.

(3) Amy and Mark (L1 WHAT, light industrial candidate)

123456789101112131415

Amy

MarkAmy

Mark

AmyMarkAmyMarkAmyMark

Okay. When you were um: (1.0) okay so you were atGromer and Zane, a cashier and customer service?Yes: : .Okay, um: . . tell me a little bit about the dutiesthat you did there.Uh: I was doing stocking, cashiering, .. uh::helping: customers: direct them to what they nee/dedOkay,an : : : d==Did you like /that job?Yeah.Why'd you lea:ve.Uh: had I had a medical uh: I had a medical lea:veand then I had personal problems plus I was lookingfor advancement. An:d things were just weren't

Getting the Job 115

161718192021222324252627282930313233343536

Amy

Mark

AmyMarkAmyMarkAmy

Mark

AmyMark

working out right for I mean . . before . . it was : :y'know. /I need \some advancement and I'm justbarely surviving from there.Okay, (.hhh) now- you said that your last day was inMay?Yeah .. I jus:t yeah I jus:t I just got um I justgot done from that job over there.When was your last /da:\y.It wa : : s two days ago.Two days ago?==Yeah.==0kay. (3.0) And you said that you are on /me/di/calleave from there?Well no I had um a : I had urn: I had a medicalproblem I had to go to the doctors . But then uh Ihad um: a: I had a medicals leave from there butthen my excuse .. is .. is: is I d' know it's just mymy boss said it's like .. y'know they had to let mego . . from that .Okay so they let you go.Yeah.

At first it appears in lines 1-5 that Amy is interested in knowing what skillsMark utilized in his job with Gromer and Zane. She does not want long,elaborate answers, however, and communicates this as follows. In response toAmy's request of Mark to tell her "a little bit about the duties that you did there"(lines 4-5), Mark begins to list some of the tasks he had, but when he slowsdown his pace with a long "an:::d," Amy takes the floor and asks a newquestion—whether he liked that job (linelO). Mark's monosyllabic answer,"yeah" (line 11), seems to suffice, as Amy again moves on to a new question,this time asking him why he left that job (line 12).

Now Mark takes a longer turn—without being interrupted—during whichhe offers four reasons for having left the job (lines 13-18). First, he states, hehad a medical leave (line 13). Second, he had personal problems (line 14).Third, he was looking for advancement (lines 14-15); here, the implication isthat he was not finding the advancement he was looking for at this particularjob. Fourth, and as a way of summarizing his previously named three reasons,Mark states in more general terms that "things ... just weren't working outright" at that job (lines 15-16). Mark then returns to his third reason, stressingthe importance of advancement, and supports that statement with a furtherelaboration that he is "just barely surviving" at that job (lines 17-18). This set ofreasons arouses Amy's suspicion enough that she then begins to verify hisemployment dates.

Amy verifies with Mark that his last day on the job was in May (lines 19-20)—the same month that this FastEmp job interview takes place. Markconfirms this, and also adds a slight amount of precision to that fact by statingthat he "just got done from that job over there" (lines 21-22). Amy, seekingmore precision, asks Mark when the last day was (line 23). Mark answers bytelling her, "It wa::s two days ago." (line 24). Suspicious again, Amy repeatsMark's statement as a question (line 25), which Mark confirms again, allowingfor no pause between their turns (line 26).

116 Kerekes

Amy's suspicion stems from the fact that, if Mark left his previous job formedical reasons only 2 days ago, chances are he should still be on medicalleave; so why is he now looking for another job? After 3 seconds of silentpondering, Amy asks for clarification by restating as a question his claim that heis on medical leave from the job he left 2 days ago (lines 27—28). Mark hedges,answers in the negative to her yes/no question (line 29), and then fumbles toprovide a better explanation for the evident incongruity. His fumbling ismanifested through repetitions (line 29: "I had um a: I had um: I had a"); pausesand fillers (lines 32—33: "my excuse .. is .. is: is I d'know it's just my my bosssaid it's like .."); and a final new (i.e., fifth) reason for having left the job, that"they had to let me go" (lines 33-34).

As far as Amy is concerned, the fifth reason offered by Mark is the onlyreason that matters, and she sums up his somewhat confusing narrative with onesentence—"Okay so they let you go" (line 35), raising her volume on "they" and"you" to indicate she now understands that the employer is responsible forMark's having left the job. Amy does not give him the benefit of the doubt byconsidering the possible legitimate reasons Mark may have been let go.Throughout the remainder of the interview, Amy focuses solely on dates andlengths of time Mark spent on various assignments, without showing anyinterest or giving Mark the opportunity to talk about what he actually did onthose jobs, what skills he has, or what other characteristics and qualifications hehas.

Predictably, Mark fails his job interview, for reasons which are identified byAmy in her assessment of him:

I just didn't get a good feeling from him. ... I think that he was trying to um, hewas trying to cover up the fact that he had a lot of employment gaps until Idelved in deeper. And so I didn't like that he was trying to be misleading.

LaQuita. In contrast to Mark's outcome, we see in the following excerptfrom LaQuita's job interview that it is possible to provide an acceptable reasonfor having left a previous job, as well as to do so in such a way as to highlightsome of the candidate's positive characteristics which might impress the staffingsupervisor favorably. LaQuita is an African American NS in her mid-twenties,applying for a clerical position. In the following excerpt, Amy asks LaQuitaabout the size of the company for which she had previously worked, and thenprobes for the reason that she left this job.

(4) Amy and LaQuita (NS, clerical candidate)

12345678

AmyLaQuita

AmyLaQuitaAmy

Okay, okay \great /how big of \a /com\pany is it,Oh gosh um .. ((laughs)) (.hhh) um I would sayimmediately, . . yeah 's 'bout, /probly \like, problylike three hundred or so==/Oh \okay so there's .. \hund/reds,==Yeah==0kay. Great. And- what was the reason that you/left this position?

Getting the Job 117

9 LaQuita Basically um I'm on an indefinite leave um I have10 the um, I have the oppor/tunity to come \back which11 is great had a big goin away party, /only /I'm I'm12 think\ing I'm lookin' for somethin a little bit13 mor:e stable as well as I'm goin' to school in the14 evenings=15 Amy =mhmm=16 LaQuita =so I'm not really um, in terms of sure if I'm gonna17 y'know go do that on a full time basis as of18 yet,[that's]19 Amy [(.hhh)] What was instable about this job?20 LaQuita /Well, in terms of, \of the instability, I just21 think we had a lot of .. comings and goings and when22 you have a [manager]23 Amy [lot of turnover?]24 La Quita /Yeah, \I think and just in terms of .. um .. the25 peopl:e .. /I mean, \it's a good envir- I mean goo:d26 company, oh in terms of the jo:b description that27 I've gone on but /I think \in terms of stability,28 y'know you kind of want a manager you kind of want29 supervisors you kind of want a little more30 structure=31 Amy =Mhmm=32 LaQuita =So I think um .. that that's definitely my reason33 for taking a leave.34 Amy Okay.

In lines 1-6, Amy conies to understand that LaQuita worked for a relativelylarge company. Amy then asks her for the reason she left the position (lines 7-8). LaQuita's response to this query highlights four positive characteristics:First, the fact that she is on indefinite leave and can return to the job if/when shewants to indicates that the decision to leave the job was hers and not heremployer's. Second, her colleagues like her so much that they gave her a going-away party, again implying that they would like to have her back, which in turnimplies that she is a desirable employee. Third, the reason LaQuita offers forhaving left the job, and for applying for a different job, is that she seeks morestability than was offered at her previous job (lines 12-13). Stability, in general,is a frequently-occurring term in the language of staffing supervisors, who seekstable job candidates to hold stable positions. LaQuita's use of the word thushelps to establish co-membership with Amy, and connotes positivecharacteristics; she also subsequently explains, per Amy's request (line 19), thatwhat she means by instable is a fast rate of turnover among her supervisors(lines 20-30). LaQuita's justification for needing a more stable workenvironment is that, in addition to wanting to work full time, she is going toschool in the evenings (lines 13-14). The pursuit of further education in theculture of FastEmp implies ambition and a desire for self-improvement, alsotraits sought by the staffing supervisors in their potential employees. Thus wesee that LaQuita has gone far beyond giving a literal answer to Amy's questionabout why she left her previous job: She demonstrates desirable traits such asbeing a "stable" employee, being well-liked by her colleagues, and beinghardworking.

Amy's assessment of LaQuita provides further evidence of the success ofher job interview. Amy identifies as of interest not only job skills LaQuita

118 Kerekes

possesses, but also, in particular, her self-presentation and interactional style,which Amy sees as professional and appropriate for a job with FastEmp:

She's good, I mean she, y'know obviously she has a good professionalappearance. . . . She came off very well with her presentation. . . . She knowsgood things to say. She knows responses such as what are your strengths and shesaid immediately communication and flexibility. Y'know some people . . . kindof stop and pause and hem and haw and not know what to say but I mean she ispoised and she is definitely sophisticated I think in her business sense so I thinkthat is definitely a strength for her.

AFTER THE EVENT

Differential Treatment of Candidates

At the conclusion of each job interview, Erin and Carol generally have thecandidates complete federal forms for tax and employability purposes, whichqualify the job candidates for work assignments without having to return toFastEmp. Amy, on the other hand, is least concerned with accommodating thejob candidates, and most concerned with efficiency. Reasoning that, if acandidate really wants the job, she or he would be willing to return to FastEmpto complete the paper work before starting the assignment, Amy generally doesnot deal with the legal documents, choosing instead to wait until the time that acandidate is assigned a job.

As with the paper work, the staffing supervisors also differ in theirapproaches to checking the job candidates' professional references. Carol ismost rigid about checking candidates' references before their job interview,feeling that, especially in light of her lack of familiarity with her new position,she must rely on these references to help her assess the candidates during theirinterviews. She also acknowledges that other factors, such as the tests thecandidates take, play into her final assessment of the candidates. Amy, on theother hand, feels comfortable making assessments of many of her candidatesbefore checking their references. In fact, again in her effort to save time, sheoften postpones checking references until she has a concrete job possibility tooffer the candidate.

Amy's standards for judging the reliability of the references varyconsiderably, depending on her overall impression of the candidate. Thecandidates are instructed on their application to provide names, titles, and phonenumbers of three "professional/work" references—"references that are notrelated to you who have knowledge of your skill level." According to Amy,candidates are to list their prior supervisors here, despite the fact that, in thework history section of their applications, there is also a space where candidatesare to list the names and phone numbers of their supervisors. The job candidatesexhibit a range of interpretations of the written instructions, however. Some listtheir supervisors, coworkers, personal references (e.g., friends), or somecombination of these. Amy's tolerance for candidates' interpretations which

Getting the Job 119

deviate from her expectations varies tremendously. Patty, for example—a whiteNS light industrial candidate—impresses Amy as nervous, not very believable,and a person who "didn't understand the things I was asking and was just givingme whatever answer came to her mind." Accordingly, Amy judges Patty'sreference choices harshly. Patty provided the names of three coworkers, one ofwhom was a close friend as she revealed to Amy during the interview: "She w-worked with me at Alamo Rictor for a long time. She's like a retired woman andwe're very close." Amy accused her, however, of providing only personalreferences, and explained to me in her debriefing interview that references mustbe supervisors and managers, as opposed to coworkers, "people that will bemore objective than somebody that's her best friend."

In contrast, Amy responds more benevolently to Liz, a white NS clericalcandidate who, similarly, does not provide the types of references Amy expects(see also Kerekes, 2003b). After Amy discovers that the first reference Liz haswritten down is a person Liz "never actually worked with," she simply asks Liz,"Okay, are any of these business references?" and then asks Liz's permission tocontact some of the supervisors whose names Liz has not provided in thereference section of her application. Liz cannot provide the phone number forone of the references Amy proposed, stating, "You can call her if you can findher number. I have no idea where it is," to which Amy responds, "I'll just callthe HR department." Amy overlooks Liz's failure to list supervisors as herreferences because Amy is impressed with her candidacy, and says of Liz, "Ithought she was good. ... I thought she was professional and serious about herjob search." Amy is similarly unbothered by the fact that Linda, a white lightindustrial job candidate, does not list former employers as references. Shesimply looks at Linda's work history and asks for Linda's permission to callsome of the supervisors she has listed in that section. In her interviews of fourother high-level white candidates for clerical positions, Amy neglects to askanything about their references during the interviews, and readily assesses themas highly employable in her debriefing interviews with me, despite the fact thatshe has not yet checked their references.

One area of agreement in the staffing supervisors' approaches to dealingwith the job candidates is their style of rejection of job candidates; theyunanimously choose to reject candidates indirectly. The staffing supervisorsgenerally tell candidates whom they do not expect to place on an assignmentthat they are still checking the candidates' references. In the case of onecandidate, Carol rationalizes this response with the following comment to me:

What I told her was that we couldn't hire her at this time. We were awaiting stillmore references. And that is somewhat true. If I get another reference thatcomes back and it's good I may change my mind.

Carol does not actually check this candidate's other references, however.The staffing supervisors never simply state to disqualified job candidates

that the candidates will not be placed on an assignment through FastEmp. EvenNguyen, a candidate whose English ability Carol deemed so weak that he did

120 Kerekes

not get to have a full job interview, was told that she would check his referencesand get back to him. In this case, Nguyen took Carol's message seriously, andhe returned to FastEmp the next day to see if she had finished checking hisreferences. Carol told him that she was still checking his references. Carol toldme, meanwhile, that if Nguyen continued to come in or call FastEmp, she wouldhave to "be more direct" with him to indicate he cannot be hired. Carol did notin fact intend to check his references, having classified him as not employableby FastEmp on the basis of his "poor command of English language." Nguyencame back later that week to check with Carol again.

In spite of Nguyen's apparent difficulty with English, his resume and testscores indicate competence in a variety of light industrial work settings. Thisraises the question of how important NNS candidates' L2 ability is as aqualification for some of the light industrial jobs. Here, again, the staffingsupervisors differ in their perspectives on how to assess the candidates. Onoccasion, a job candidate arrives at the job interview with an interpreter (usuallya family member or friend). Carol's perspective is that, since "applicants can'ttake their translators with them to the job," she cannot hire them if they need aninterpreter for the interview. In the case of other NNSs, Carol uses the question,"What are your long-term goals?" to gauge their L2 ability. Carol's experiencehas led her to conclude that "if they don't have a good command of the Englishlanguage, they don't understand that question." Amy, in contrast to Carol, takesthe approach that NNSs with low English ability can be hired for jobs at whichthe supervisors can communicate with such candidates in their L1.

Successful, Failed, and Weak Job Interviews

On the basis of assessments offered in the debriefing interviews with the staffingsupervisors, documentation of the job candidates' statuses, and transcripts of thejob interviews, each of the job candidates in this study was categorized ashaving had a successful, failed, or weak job interview.9 The success of a jobcandidate was revealed during debriefing, in the staffing supervisors' answer tothe question, "Do you think you'll be able to place this person?" The staffingsupervisor's answer to this question was an unequivocal "yes" for successful jobcandidates. The one exception to this rule was a candidate whom Carol (thestaffing supervisor who interviewed her) wanted to place, but who becamedisqualified on the basis of a negative reference from her former employer. Jobcandidates who had failed interviews were those whom the staffing supervisordid not intend to place on an assignment. Several job candidates werecategorized as weak because the staffing supervisor felt she could place them onassignments despite their interviews. That is, although they did not impress thestaffing supervisor with their interviewing behavior, the skills they displayed

9 100% intercoder reliability was established for this categorization, using approximately 10% of thedata (randomly chosen) and two coders.

Getting the Job 121

and references they provided indicated to her that they could successfully carryout a FastEmp assignment.

The relative success of each candidate depended, however, on other factorsas well, such as the current availability and/or need to fill particular workassignments, or projected needs in the near future. Some candidates who wereseen as marginal at best (especially some of the light industrial candidates), werehired or given legal forms immediately after their interviews (thereby makingthem eligible to accept an assignment immediately, without having to return toFastEmp before starting a job), because they could be used to fill job orders thathad already been placed by client companies of FastEmp.

Distribution of Outcomes

In their assessments of the job candidates, the staffing supervisors identifiedthree salient areas in which the candidates are expected to be proficient: jobskills, language use, and ability to conduct themselves skillfully in a jobinterview (interviewing competence). Many of the issues they discuss involveskills they feel can be specifically learned, practiced, and improved by jobcandidates in order to achieve better results in the future. These include theability to predict questions that will be asked and to have prepared answers forthese questions, specifically indicating character strengths such as enthusiasm,leadership or the ability to take initiative, flexibility, reliability, and trustworth-iness. Now we shall see which of the candidates tended to have greater successand which had higher failure rates, on the basis of socioeconomic variables,including the types of jobs they seek, the candidates' genders, their race, andlanguage backgrounds.

Distribution According to Job Type. Of the 24 job candidates who appliedfor clerical positions, 20 (83%) had successful interviews, one (4%) had a weakinterview, and three (13%) had failed interviews. Of the 24 job candidates whoapplied for light industrial positions, in contrast, only 11 (46%) had successfulinterviews. In other words, nearly twice as many clerical candidates as lightindustrial candidates had successful job interviews. Eight (33%) of the 24 lightindustrial candidates were not hired (i.e., they had failed interviews; thesenumbers include Nguyen, the candidate who was disqualified because of his L2ability before he could have a job interview). The remaining five (21%) were

122 Kerekes

Table 4.2. Job Interview Success Rate, by Job Type

Job type

Light industrial

Clerical

Total

Successful

11

20

31

Success rate

Weak

5

1

6

Failed

8

3

11

categorized as weak because, even though they qualified for work, theirinterviewing skills were deemed inadequate by the staffing supervisors whoassessed them. Table 4.2 illustrates the distribution of successful, weak, andfailed job interviews across light industrial and clerical job candidates.

Distribution According to Language Background and Job Type. Overall,the success rates according to language background did not differ notably, norwas there much difference in the success rate by job type. Within each of thecategories described earlier, we can see the distribution according to languagestatus and race in Table 4.3. Of the successful light industrial workers, 7 (63%)are NNSs (all people of color, since, in this study, all of the NNSs are people ofcolor), and 4 (36%) are NSs, including two African Americans and two whites.Of the light industrials who had weak or failed interviews, 7 (54%)are NNSs(again, all people of color) and 6 (45%) are NSs (three African American andone White).

Of the five NNSs who failed their interviews, two were disqualified onaccount of their inadequate L2 ability, according to the staffing supervisors'assessments (one of these was Nguyen, who did not have a complete job inter-

Table 4.3. Job Interview Success Rate, by LI and Job Type

Job type Language Success rate (by background)status

Light NNSindustrial NS

Clerical NNSNS

Total

Successful

7 (4A, 2L, 1P)4 (2B, 2W)

8 (6A, 2L)12(3B,1L,8W)

31

Weak

3(3L)2 (1B, 1L)

1(1P)0

6

Failed

4 (2A, 2L)4 (3B, 1W)

1(1A)2(2W)

11

Note. A = Asian; B = Black/African American; L = Latino; P = Pacific Islander; W = White/Euro-pean American. (Indicates background information)

Getting the Job 123

view). In addition, staffing supervisors mentioned minor concerns about the L2ability of five of the NNSs who, nevertheless, had successful job interviews.

They were also concerned about the L2 ability of one job candidate whohad a weak interview, but who was subsequently placed successfully on anassignment.

Distribution According to Gender and Job Type. For both the clerical andlight industrial job types, more of the female candidates were successful than themale candidates. Of the 24 female job candidates, 20 (83%) had successfulinterviews, whereas only 11 (46%) of the 24 male candidates had successfulinterviews (Table 4.4); nearly twice as many women as men had successfulinterviews.

In terms of gender and job type, the least successful group of candidates inthis study were the male light industrial candidates. Of the 14 male lightindustrial candidates, only four had successful interviews. Ten of the 14 malelight industrial candidates, or 71%, had weak or failed interviews. Not only didthese candidates (the male light industrials) differ from the staffing supervisorsin gender, but also in educational and socioeconomic backgrounds; the staffingsupervisors, unlike the male light industrial job candidates they interviewed, hadall received at least 2 years of higher education and came from middle-classbackgrounds. Many of the light industrial candidates came from self-describedworking-class backgrounds. In other words, the candidates who had the least incommon with their interviewers had the highest failure rate. In terms of race, wesee that, of the 10 who failed or had weak interviews, all of them (100%) weremen of color (two Asians, three African Americans, and five Latinos).

Distribution According to Interviewer. Finally, there were alsodifferences in success rates of job interviews depending on which staffing

Table 4.4. Job Interview Success Rate, by Gender, Job Type, and Ethnic Background

Job type

Lightindustrial

Clerical

Total

Gender

MaleFemale

Male

Female

MaleFemale

Success rate (by background)

Successful

7 (2A, 2B, 1L, 1P, 1W)4 (2A, 1L, 1W)

13 (3A, 3B, 3L, 4W)

7 (2 A, 1L, 4W)

20 (5A, SB, 4L, 1P, 5W)11 (4A, 2L, 5W)

Weak

05 (1B, 4L)

0

1(1P)

06 (1B, 4L, IP)

Failed

3 (1B, 1L, 1W)5 (2A, 2B, 1L)

1(1W)

2 (1A, 1W)

4 (1B, 1L, 2W)7 (3A,2B, 1L, 1W)

Note. A = Asian; B = Black/African American; L = Latino; P = Pacific Islander; W = White/Euro-pean American. (Indicates background information)

124 Kerekes

Table 4.5. Job Interview Success Rate, by Staffing Supervisor, and Job Type

Job type Success rate in number of interviews

Successful

Light industrial

Clerical

Total

AmyCarolErinRenata

AmyCarolErinRenata

AmyCarolErinRenata

2711

9470

111181

Weak

AmyCarolErinRenata

AmyCarolErinRenata

AmyCarolErinRenata

0320

0010

0330

Failed

AmyCarolErinRenata

AmyCarolErinRenata

AmyCarolErinRenata

6200

1020

7220

supervisor conducted the interview (Table 4.5). Amy, who conducted thehighest number of job interviews in this study (totaling 18) had the highest totalnumber of failed interviews (7, or 39% of her interviews) and no weakinterviews. Although Erin and Carol each had fewer total interviews, eachhaving three weak and two failed interviews, the ratio of success to failure wassimilar across all three interviewers. (Erin's weak and failed interviewsamounted to 38% of her inter-views, and Carol's weak and failed interviewsamounted to 31% of her interviews.) The rate of successful interviews was 61%for Amy, 62% for Erin, and 69% for Carol.

Amy's interviewing style reflects her relative lack of tolerance, incomparison to Erin and Carol, for inconsistencies in the job candidates' self-presentation and answers to her questions. This may explain why she failedmore of the candidates and made a clearer distinction between successful andfailed candidates (with no marginal, weak candidates) than the other staffingsupervisors.

DISCUSSION

The findings from this study show a connection between the job candidates'race, gender, and the success of the job interview. The fact that the three mainstaffing supervisors in this study are European American middle-class women isnot coincidental (most staffing supervisors in this industry are). Beyond thesebiographical similarities, they also share goals and values: All of them arebetween the ages of 25 and 30; they have similar ambitions; their positions asstaffing supervisors are stepping stones to attaining other career goals; they arepart of corporate America; they value higher education, and plan to continuetheir own educations in the near future.

The group of job candidates whose failure rate is highest consists of maleapplicants for light industrial positions, and they are all minority races (two

Getting the Job 125

Asians, two African Americans, and one Latino; see Table 4.4). They have lesseducation and fewer plans to continue their education than the staffingsupervisors have; the type of work in which they are interested is vastly differentfrom the world of work with which the staffing supervisors are familiar; they areof the opposite gender from the staffing supervisors; and, in general, there islittle overlap between their world and that of the staffing supervisors.

The female job candidates in this study had greater success than the malecandidates, either because they were better able to figure out how to act in anacceptable manner, or because doing so came more naturally to them. Followingthe argument that similar backgrounds breed similar Discourses (Gee, 1996),10

immediately in favor of the female job candidates is the fact that they are of thesame gender as the staffing supervisors. One of the characteristics of about onethird of the successful job interviews was the existence of rapport-building chit-chat during the interviews. Incidental topics which were discussed includedcommon acquaintances, raising children, and getting married. Involvement ofthis nature has been associated by some scholars with a female conversationalstyle (Tannen, 1990). While, in this study, four of the nine candidates whoemployed chit-chat (demonstrating comembership) with the staffing supervisorduring the job interview were men, all nine who did so had successfulinterviews. One could argue that, whether or not characteristic of femalespeakers, chit-chat is certainly one of the characteristics of the interactionalstyles of the female staffing supervisors in this study. By engaging in informalchit-chat during the interview, the job candidates find common ground with thestaffing supervisors' discourse styles, and vice versa. Because of lack of datafrom male staffing supervisors (of whom there are none at FastEmp), we cannotsay that the chit-chat phenomenon is gender-specific, but we can say that itcharacterizes some successful gatekeeping encounters.

In the case of FastEmp, a common discourse style shared by theinterlocutors more likely results in a positive interaction than when theinterlocutors' discourse styles are vastly different from one another; the staffingsupervisors represent a higher prestige discourse style than their light industrialjob candidate. The negatively evaluated differences between their discoursestyle and that of the light industrial candidates result in less favorableinteractions, because the staffing supervisors react more positively to discoursestyles similar to their own. One's discourse style reflects one's familybackground, educational background, and class. In this study, the staffingsupervisors reacted more favorably to those candidates whose backgroundsmore closely resembled their own.

10 Discourses (spelled with a capital "D"), as defined by Gee (1996), are ways of displayingmembership in particular groups, through a display of values and beliefs related to one's socialidentity. These are expressed through a speaker's ways of acting, talking, writing, and interactingwith others, and can involve not only the language itself, but also gestures, physical appearance, andeven other props which communicate one's identity (e.g., a symphony musician who carries herinstrument with her, or a job applicant's briefcase).

126 Kerekes

Before, During, or After?

We have seen a number of influences on the staffing supervisor's evaluation of ajob candidate, some of which operate before the staffing supervisor meets thecandidate face-to-face, others which operate immediately upon their firstmeeting in person, and yet others which come into play during the actual jobinterview. Beyond these, however, still another source influences the staffingsupervisors' assessments of the candidates: the greater context of theinteraction, or the "pretext" (Hinnenkamp, 1992, cited in Meeuwis, 1994). Thestaffing supervisors' preconceived notions of job candidate types, based ongender, race, education, and the type of job for which they are applying, promptthose supervisors to have certain expectations of the individual job candidatesand to interact with them accordingly. Similarly, the job candidates'predetermined ideas of FastEmp itself, and of the employment agency jobinterview setting in general, influence those candidates' behavior and reactionsduring the gatekeeping encounters.

Among the staffing supervisors, the label "light industrial" conjures upimages of candidates who are "casual," "lax," "not professional," and who have"weak" speaking skills (e.g., "poor grammar").11 Light industrial candidatesare, therefore, at a disadvantage in comparison to their clerical counterparts,whom the staffing supervisors do not automatically expect to be below parregarding professionalism. Demographic trends at FastEmp are such that, if acandidate is both male and a racial minority, he is more likely to be applying forlight industrial work than for clerical work. As a light industrial candidate, healready has points against him before he commences his job interview. Negativeexpectations of light industrial candidates get transferred to more general,negative expectations of African American, Latino, and Asian male candidates.These negative expectations cause staffing supervisors to be intolerant ofcandidates' deviations from ideal job interviewing conduct. While thesecandidates may be able to compensate for the initial negative judgments theyface, they must work hard to do so. Clerical candidates and those candidateswhom the staffing supervisors generally label as professional, on the other hand,are more likely to be given the benefit of the doubt if and when they deviatefrom expected job interviewing conduct.

The interviewer constrains the possible discourse types that are used—thoseof the dominant cultural group (Fairclough, 1989). Participants' preconceivedideas about their interlocutors, or about the culture/people they believe theirinterlocutors represent, are influenced by their ideological presuppositions(Foucault, 1972; Shea, 1994) and related discourses (Gee, 1996). Interlocutors'power differences on the basis of skin color, gender, and lifestyle can serve tolegitimize and reproduce boundaries between members of different groups;

11 These descriptors are in quotes because the words are taken from the staffing supervisors'assessments of various light industrial job candidates.

Getting the Job 127

these are manifested in their verbal interactions (Bernstein, 1996; Fairclough,1989). Therefore, even when saying essentially the same thing, if speakers lookdifferent (e.g., different races or different genders) or speak differently (e.g., aBritish accent versus an Indian accent)—that is, if different expectations are heldof the speakers (Tannen, 1993a, 1993b)—they are likely to be interpreteddifferently as well (Gumperz, 1982b) This is often to the detriment of speakersof low-prestige varieties of English (including normative and nonstandardvarieties). Different interpretations can lead to misunderstandings which are notnecessarily recognized or acknowledged: "The absence of superficial problemsof meaning exchange is no indication of the absence of miscommunication atdeeper levels" (Coupland, Wiemann, & Giles, 1991, pp. 6-7).

In addition to preconceived biases or stereotypes the staffing supervisorsmay have of the job candidates, the amount of data from which the staffingsupervisor forms her impression of the candidate before commencing the jobinterview varies tremendously, and is often determined by chance occurrencesand/or chance timing of the occurrences. Depending on the current demand fromclients for temporary employees, how many staffing supervisors are in theoffice, the time of day, and numerous other factors, the staffing supervisors willbe more or less familiar with a candidate's resume, application, test scores, andreferences before the interview commences. However accidental the set offactors is which contributes to the preinterview impression, it clearly affects theinteractions that subsequently occur during the interview. Add to theseimpressions the staffing supervisor's immediate impression upon seeing the jobcandidate—Is the job candidate on time? What is she or he wearing? Howprofessional does this candidate look?—and the staffing supervisor likely hasquite a distinct idea about this candidate's eligibility for work before the jobinterview has even begun.

Can a Judgment Be Reversed?

Job candidates who are received by the staffing supervisors with negativeexpectations are not automatically doomed to fail their interviews, but they mustbe able to implement compensatory strategies in order to (re)gain theirinterlocutors' trust, against the odds. One light industrial candidate—a Mexicanmale in his 20s—impressed the staffing supervisors when he arrived at hisinterview dressed in slacks and a white dress shirt. Mild-mannered and clean-cut, he impressed them with his "professionalism." Stated another way,although he was applying for light industrial work, he looked and acted like theirprototypical, successful, clerical candidate; his appearance thus worked in hisfavor.

In the case of a job candidate who is distrusted by a staffing supervisorbecause of gaps in her or his work history, the staffing supervisor is likely tofocus her questions on the precise dates and locations of the candidate'sprevious assignments, rather than asking the candidate about her or his strengthsand experiences in former positions. The way the candidate chooses to answer

128 Kerekes

her questions can serve to establish either legitimacy or illegitimacy. Forexample, in the case of Martin, an African-American NS light industrialcandidate, spending 4 months looking for work was not viewed as a legitimatereason for gaps in his work history; in contrast, taking time off to rest and relaxat Lake Tahoe was considered legitimate, as in the case of Peter, a White NSclerical candidate (Kerekes, 2003a, 2003b). Initially, Peter aroused Amy'ssuspicion on account of his written application, which indicated gaps in his workhistory; he was able to compensate for the flaw in his application by giving Amyan acceptable reason for the gap, however.

Another way the candidates establish legitimacy is by knowing when toelaborate on their answers. The successful—and trusted—job candidatesvolunteered more information about themselves and their circumstances than thedistrusted candidates, who often answered the staffing supervisors' questionsliterally, without adding any information other than that which was specificallysolicited. Similar results were found in studies of Oral Proficiency Interviews ofNNSs, in which candidates who provided minimal instead of elaborated answersrisked underrepresenting their L2 ability and thus receiving lower proficiencyratings (Ross, 1998; Young & Halleck, 1998). Even within the realm ofelaboration, however, the candidates must know what sort of elaboration islegitimate in the gatekeeping encounter.

Going into great details about one's technical expertise in an area withwhich the staffing supervisor is not very familiar, for example, is not anappropriate kind of elaboration. Because, especially in the case of lightindustrial positions, the staffing supervisors do not possess the expertise to beable to assess the candidates' field-specific job qualifications, they emphasizethe skills they are more able to evaluate: interpersonal skills, generalprofessionalism, and trustworthiness. These criteria have been chosen as thefocal areas for consideration in this mediated interviewing situation notnecessarily because they are the most crucial, but because they are mostcompatible with the staffing supervisors' areas of expertise. The avoidance ofqualities the staffing supervi-sors cannot judge well (e.g., a light industrialcandidate's familiarity with various tools and pieces of equipment) and attentionto those a staffing supervisor feels competent to judge involves a shift fromprofessional qualifications to personal qualities.

Are Successful Interviews Successful?

In this chapter, we have seen that, while clerical candidates had a higher successrate than their light industrial candidates, and female candidates were moresuccessful than their male counterparts, there was no obvious advantage to beingeither a NS or NNS of English; neither was there an advantage to having aparticularly high level of L2 ability, for the NNSs. Successful encounters wererelated to the ability to demonstrate desirable employee characteristics, which

Getting the Job 129

proved equally possible for job candidates with a variety of languagebackgrounds.

What we have not seen is how accurate, or inaccurate, the staffingsupervisors' judgments of the job candidates proved to be. What must still bequestioned is just how effective the interviewing process is for determining whowill succeed and who will not, on the job. Since data collection took place overseveral months, it was not possible to follow up equally on the progress of all ofthe candidates. What I did find, however, was that, at the time of my departure,of the 37 job candidates who had originally been assessed by the staffingsupervisors as employable, eight had since been fired from FastEmp (for poorjob performance or for not showing up for a job assignment), and another sixhad never been offered any work opportunities through FastEmp, despite theireligibility. Of the remaining 23 candidates eligible for employment throughFastEmp, 11 were currently on assignments; 4 were not working but wereavailable to be placed on assignments; and 8 had become inactive in thedatabase, meaning that they were no longer being considered for assignments(this may have been because they had found full-time employment, were nolonger interested in working for FastEmp, or were not any longer seen assuitable for FastEmp assignments). A high number of job candidates who hadsuccessful job interviews did not end up being successful on their assignments;we can only wonder how many of the candidates who had failed interviews mayhave ended up working out well on the job, had they been given the chance.Although it is in the interest of job candidates to learn how best to impress thestaffing supervisors with their interviewing competence, it is in the interest ofFastEmp staff to consider how to raise the success rate of their interviewingprocess— taking success to mean an accurate evaluation and resultingplacement (or nonplacement) of the job candidates.

REFERENCES

Akinnaso, F. N., & Ajirotutu, C. S. (1982). Performance and ethnic style in job interviews. In J.J.Gumperz (Ed.), Language and social identity (pp. 119-144). New York: CambridgeUniversity Press.

Bernstein, B. (1996). Pedagogy, symbolic control, and identity. London: Taylor & Francis.Coupland, N., Wiemann, J. M., & Giles, H. (1991). Talk as "problem" and communication as

"miscommunication": An integrative analysis. In J. Wiemann (Ed.), Miscommunication" andproblematic talk (pp. 1-17). Newbury Park, CA: Sage.

Denzin, N. K., & Lincoln, Y. S. (1994). Introduction: Entering the field of qualitative research. InY. S. Lincoln (Ed.), Handbook of qualitative research (pp. 1-17). Thousand Oaks, CA: Sage.

Erickson, F., & Shultz, J. (1982). The counselor as gatekeeper: Social interaction in interviews. NewYork: Academic Press.

Fairclough, N. (1989). Language and power. London: Longman.Foucault, M. (1970). The order of things: An archaeology of the human sciences. New York:

Random House.Gee, J. P. (1996). Social linguistics and literacies: Ideology in discourses (2nd ed.). London: The

Falmer Press.Gumperz, J. J. (Ed.). (1982). Language and social identity. Cambridge, England: Cambridge

University Press.

130 Kerekes

Gumperz, J. J. (1992). Interviewing in intercultural situations. In J. Heritage (Ed.), Talk at work:Interaction in institutional settings (Vol. 8, pp. 302-327). Cambridge, England: CambridgeUniversity Press.

Holstein, J. A., & Gubrium, J. F. (1995). The active interview (Vol. 37). Thousand Oaks, CA: Sage.Kerekes, J. (2001). The co-construction of a successful gatekeeping encounter: Strategies of

linguistically diverse speakers. Unpublished doctoral dissertation., Stanford University,Stanford.

Kerekes, J. (2003a). Distrust: A determining factor in the outcomes of gatekeeping encounters. In S.Ross (Ed.), Misunderstanding in social life, London: Longman/Pearson.

Kerekes, J. (2003b). Winning an interviewer's trust in a gatekeeping encounter. Manuscriptsubmitted for publication.

Meeuwis, M., & Sarangi, S. (1994). Perspectives on intercultural communication: A critical reading.Pragmatics, 4(3), 309-313.

Ross, S. (1998). Divergent frame interpretations in language proficiency interview interaction. In A.W. He (Ed.), Talking and testing: Discourse approaches to the assessment of oral proficiency(Vol. 14, pp. 333-353). Amsterdam: John Benjamins.

Schiffrin, D. (1994). Approaches to discourse. Cambridge, MA: Blackwell.Shea, D. P. (1994). Perspective and production: Structuring conversational participation across

cultural borders. Pragmatics, 4(3), 357-389.Tannen, D. (1990). Gender differences in topical coherence: Creating involvement in best friends'

talk. Discourse processes: A multidisciplinary journal, 13(1), 73-90.Tannen, D. (1993a). The relativity of linguistic strategies: Rethinking power and solidarity in gender

and dominance. In D. Tannen (Ed.), Gender and conversational interaction (pp. 165-188).Oxford, England: Oxford University Press.

Tannen, D. (1993b). What's in a frame? Surface evidence for underlying expectations. In D. Tannen(Ed.), Framing in discourse (pp. 14-56). New York: Oxford University Press.

Young, R., & Halleck, G. B. (1998). "Let them eat cake!" or how to avoid losing your head in cross-cultural conversations. In A. W. He (Ed.), Talking and testing: Discourse approaches to theassessment of oral proficiency (pp.355-382). Amsterdam: Benjamins.

Young, R. F., & He, A. W. (1998). Language Proficiency Interviews: A Discourse Approach. In A.W. He (Ed.), Talking and testing: Discourse approaches to the assessment of oral proficiency(pp. 1-24). Amsterdam: John Benjamins.

Getting the Job 131

APPENDIX

Transcription Conventions and Illustrations

[ ]=

(1)

-

,?7

/

\!

italics

CAPITALS0

(hhh)

(.hhh)

(text)

((phenomenon) )

( )

(x x)

{ ( (phenomenon) ) text}

overlap

latching

pause less than 0.5 second

pause greater than 0.5 second and lessthan 1 second

timed pause (in seconds)

elongation

cut-off

final falling tone

slight rise

final rising tone

weaker rising tone

higher pitch in following syllable (s)

lower pitch in following syllable (s)

animated tone

slightly louder volume

much louder volume

softer volume

audible aspiration (out-breath)

audible inhalation (in-breath)

transcriptionist doubt; a good guessat an unclear segment

vocal or nonvocal, nonlexicalphenomenon which interrupts lexicalstretch

unintelligible speech

unintelligible speech with a goodguess at the number of syllablesindicated by number of x's)

vocal or nonvocal, nonlexicalphenomenon that co-occurs withlexical segment indicated betweencurly brackets .

5

Discourse Strategies in the Context ofCrosscultural Institutional Talk:

Uncovering Interlanguage Pragmatics in theUniversity Classroom

Catherine Evans DaviesThe University of Alabama

Andrea E. TylerGeorgetown University

A number of researchers have noted that the field of interlanguage pragmatics(ILP) has focused primarily on comparisons of the L2 learners' production ofspeech acts to those of native speakers (e.g., Bardovi-Harlig, 1999; Kasper,1992; Kasper & Schmidt, 1996). Kasper and Schmidt (1996) suggested that ILPstudies have remained comparative in nature, at least in part, because they haveprimarily been modeled on crosscultural pragmatic analyses in which thepurpose is to examine the similarities and differences between two distinctspeech communities, rather than to understand the L2 learner's development oftarget-like pragmatic competence. Moreover, the majority of these studies haverelied on data elicited from various forms of discourse completion tasks (DCT).The weaknesses and limitations of this method of data collection are well known(See Introduction, this volume). Thus, the foci of interlanguage pragmatics havebeen highly limited in methodology of analysis (comparative), data (elicitedresponses to DCTs), and theoretical framework (speech act theory). The purposeof the present chapter is to begin to address all three of these limitations. Theprimary goal is to demonstrate that naturally occurring institutional talk is a richsource of data that, when analyzed from a triangulated, interactionalsociolinguistic perspective, can yield significant insights into the variousdiscourse dimensions which contribute to interlanguage pragmatics.

We present an example of naturally occurring institutional talk thatillustrates interlanguage pragmatics in the discourse of an advanced L2 learner.Rather than simply comparing the L2 learner's production to that of nativespeakers, the analysis considers the norms of both the speaker's L1/C1 (firstlanguage/first culture) and the target language/culture. The analysis reveals thatthe L2 learner uses a number of discourse strategies that appear not to berepresentative of the norms of either his native language, Korean, or the targetlanguage, U.S. English. Interfaces between two cultural systems are whatKramsch (1993) calls "third places." Like instances of interlanguage in the areasof syntax or phonology, this finding constitutes evidence of a stage in thedevelopment of the L2 learner's interlanguage pragmatics that cannot be

133

134 Davies and Tyler

straightforwardly explained as simple interference from the L1 or negativelanguage transfer. Additionally, we present an analytical framework for un-covering such instances of interlanguage pragmatics that is grounded in inter-actional sociolinguistics (Gumperz, 1982, 1992) and relies on triangulation ofthe data through video playback with the participants and ethnographicinterviews with members of the participants' speech communities, as well as theanalysts' detailed examination of the discourse.

Previous Research on Which This Chapter Builds

The present analysis builds on Tyler and Davies (1990) which examined aportion of this same extended interaction. That analysis focused on a discussionof grades and found mismatches in discourse management strategies, schemas,interpretive frame, interpretation of participant roles, and discourse structuringcues in the linguistic code. In that paper, we offered a working hypothesisconcerning a crosscultural difference in expectations for discourse-levelorganization patterns and discourse management styles in discussions of gradesbetween teachers and students. We hypothesized that American teachers wouldtend to adopt a "deductive/assertive" strategy when asked to explain why astudent received a low grade, probably starting with an overall summary of thestudent's execution of the assignment or with the point the instructor consideredmost important. In contrast, Korean teachers would tend to use an "inductive/collaborative" strategy by which they introduced a number of small problemswith which the student would be expected to agree and gradually move toarticulation of the major problem. By analogy with written genres, the Americanwould be more likely to begin with a thematic statement or topic sentence andthen back it up; the Korean would be more likely to save the topic sentence forthe end and rely on the listener to make the connections along the way.

We also noticed, but did not pursue in the earlier analysis, an embeddeddiscussion concerning possible cheating on the part of the student. We wereparticularly struck by the apparent shift in discourse strategy to a"deductive/assertive" style by the Korean teacher in this section of theinteraction. As native speakers of English, we were frankly surprised by thedirectness and deductive rhetorical organization the ITA used in this section ofthe interaction and felt strongly that it violated norms of the U.S. classroom. InTyler and Davies (1990) we suggested that the topic under discussion might be asignificant variable affecting discourse management strategy. In particular, wehypothesized that in a discussion of cheating, the norm for an American teacherwould be the "inductive/ collaborative" style while the norm for a Koreanteacher might be the "deductive/assertive" style. Thus, our initial hypothesis wasthat this apparently problematic discourse management strategy was the result ofnegative transfer from the L1/C1. The present analysis, which includes addition-al data from in-depth interviews with representative members of the relevantspeech communities, reveals that the issue is not so simple.

Discourse Strategies 135

An Interactional Sociolinguistic Approach

In contrast to Conversation Analysis (Markee, 2000), the framework we adoptrejects the notion that a full understanding of a crosscultural interaction can beestablished solely from the text. This is because an interactional sociolinguisticperspective accepts as fundamental the postulate that the lexical items and thesyntax in which they occur always underdetermine the rich interpretationassigned to any interaction (Green, 1989; Grice, 1975; Gumperz, 1982).Analysis of the text alone is insufficient for uncovering participants' inferences(which are by definition implicit) that are crucial elements of the interpretationand the emergent context. Gumperz points out that every language/culturedevelops a set of implicit implicatures which are associated with constellationsof features at different levels of linguistic organization (prosodic, phonetic,lexical, grammatical, pragmatic, etc.) that constitute what he calls contextual-ization cues. These cues channel interpretive processes.

Because much of the interpretation by an interlocutor stems from implicit,speech-community-specific, contextualization cues, the analyst cannot reliablyestablish the interlocutors' interpretation of the ongoing crosslinguistic/crosscultural interaction nor the motivation for an interlocutor's contribution byexamining only the text. Triangulation involving participants' commentary andthe commentary of representative members of the relevant speech communitiesis also essential. The triangulated analysis evolves out of an interweaving ofcareful attention to the discourse (including paralinguistic elements),commentary by the participants, which is elicited through video play back, andcommentary from representative members of the relevant speech communities.The analyst can independently identify specific aspects of the discourse whichreveal tensions or difficulties through careful attention to rhythm, pausing,stress, interlocutor's utterances, discourse management strategies, bodylanguage, etc. (Erickson & Shultz, 1982). These can form the basis for guidedvideo playback, but the analysis does not fully emerge until all three sources ofdata have been integrated.

Our approach relates to at least two established traditions within SLA in thestudy of classroom discourse (Ellis, 1994). One is the tradition of theethnography of speaking and communication (Hymes, 1974; Saville-Troike,1989). Ethnographic dimensions of our approach are the focus on context, theuse of video- and audiotaping, participant observation, the elicitation ofinterpretations by members of the appropriate speech community, and triangu-lation of methods. The other related approach is ethnographicaHy orienteddiscourse analysis that seeks to understand necessary sociocultural backgroundknowledge on which interpretations are based. We also seek evidence fromwithin the discourse itself of how interlocutors are reacting to each other,positioning'themselves, and interpreting the ongoing interaction.

Discourse analysis from an interactional sociolinguistic perspectiveconsiders simultaneously occurring, multiple layers of the discourse as it occurswithin a situated context (Duranti & Goodwin, 1992; Gumperz, 1982, 1992).


Contextualization cues can arise from any of these facets of the interaction. Thesituated context provides important norms of expectation for multiple aspects ofthe interaction. For instance, the situated context of the U.S. universityclassroom provides norms for the participant roles of instructor and student. Forspace reasons, we will limit the present analysis by placing primary focus ondiscourse management strategies. We will argue that the discourse produced bythe L2 learner in this exchange represents an amalgam of the learner's L1pragmatics, a partial understanding of the target language pragmatics, and theparticular restrictions and resources of the institutional setting in which theinteraction occurred.

Discourse management strategies are purposeful linguistic maneuvers thatprovide both an overall ordering to the shape of an extended stretch of discourse,as well as more local stratagems for accomplishing a particular communicativegoal in spoken interaction. Specific speech communities appear to privilegeparticular discourse management strategies such that they often achieve thestatus of cultural schemas that have an ideological component. They are alsoshaped by more specific institutional cultures and local dynamics of theemergent context. Within SLA and applied linguistics, perhaps the closestaffinity would be with work in "contrastive rhetoric" (e.g., Connor, 1996; Hinds,1987; Kaplan, 1966), which attempts to find prototypical patterns of exposition,persuasion, or argumentation associated with different cultural traditions. Workin contrastive rhetoric, however, primarily deals with written language.Discourse management strategies1 are part of ordinary everyday interaction, thusthey occupy another realm in which the prototypical schemas learned as part ofthe socialization process are typically not available to consciousness of thespeaker.2

Bardovi-Harlig (1999) has also pointed out that interlanguage pragmaticshas tended to try to artificially separate grammatical from pragmaticcompetence. The interactional sociolinguistic approach to discourse adopted

1 Our use of the word "strategy" differs from those found in the mainstream SLA literature (cf. Ellis,1994; Faerch & Kasper, 1983). There is one tradition of the "learning strategy," with relation tospeakers' development of interlanguage (Dulay & Burt, 1975), which has received comprehensivetreatment by Oxford (1990). The other tradition is that of the "communication strategy," traced by S.Pit Corder (1983) to Selinker's (1972) "strategies of communication" to account for learner errors inspontaneous speech. Canale and Swain (1980) accordingly defined "strategic competence" as theability to cope with problems in communication, and thus this use of "strategy," arising out of erroranalysis, has been defined specifically as a response to a breakdown in communication. Theinstantiation of this use of "strategy" is found in taxonomies such as Haastrup and Phillipson (1983),and Bialystok (1983): e.g., generalization, word coinage, restructuring, paraphrase, language switch,foreignizing native language terms, transliteration, semantic contiguity, description. All of these"strategies" appear to be local and specific from a discourse perspective, oriented ,to a problem in aparticular part of an utterance. Perhaps Ellis's (1999, p. 251) expression of "doubts regarding thecentrality afforded to discourse repair in mainstream SLA, suggesting that its importance for L2acquisition may have been overestimated," may liberate the word "strategy" for new uses.2 Some work on speech acts within interlanguage pragmatics (e.g., Trosborg, 1987), to the extentthat the analysis talks in terms of "strategies" for achieving, for example, an apology, is shiftingmore toward the use that we are adopting.


here is more holistic. All levels of linguistic organization are potentiallysignificant in signaling communicative intent through contextualization cues.Within this framework, pragmatic competence cannot effectively be analyzedseparately from grammatical competence; in fact, it may be the case that"pragmatics provides the most general support system for mastery of the formalaspects of language" (Bruner, 1981, p. 32). Taking an even broader perspective,the notion of "discourse strategy" allows us to move in the direction of theconcept of "interactional competence" (Hall, 1995; Kramsch, 1986) at thediscourse level.

THE DATA

The primary data under consideration is the video recording of a naturallyoccurring interaction from a routine observation in an American universityclassroom between an International Teaching Assistant (ITA) from Korea andan American undergraduate. At the time of the original data collection bothauthors were faculty within the institution, and Tyler directed a program forimproving the oral communication skills of International Teaching Assistants.For discussion of routine observations and video playback sessions, see Tylerand Davies (1990), Davies et al. (1989), Davies and Tyler (1994), Tyler (1994).We consider the ITA an advanced English learner based on his TOEFL score of620 and his SPEAK score of 240. The SPEAK test is the institutionally-administered version of the Test of Spoken English produced by the EducationalTesting Service. Two additional sources of data are verbal reports by the ITAduring several video play back sessions and interviews with representativemembers of the relevant speech communities. (See more discussion of"representative members of the relevant speech communities" below when wetalk about American and Korean interviewees.)

The physical setting of the interaction is a physics lab. Students sit inassigned pairs at lab tables. There is an instructor's desk at the front of the room.At the beginning of the lab session, while the students were taking a quiz, theITA had handed back the graded homework and a quiz from the previous week.At the top of one of the quizzes the ITA had written: "Please do not cheat." Thatstudent's lab partner was absent and so did not receive her quiz, which had thesame warning at the top. As the other students started their experiments, thestudent in question came to the front of the classroom to ask about the grades hereceived. The discussion began with the points received on the homework; inthis stretch of the interaction the ITA used an inductive/collaborative discoursemanagement strategy. In the 1990 paper, we argued that this strategy wasunexpected for the American undergraduate and contributed to the escalatingmiscommunication in which the student repeatedly challenged the ITA'sexplanations. After a lengthy and rather testy exchange about the homeworkwhich ended in the student acquiescing to the ITA's explanation, the studentpulled out his quiz and asked why the ITA had written "Please do not cheat" atthe top.


To provide the conversational context of the discussion of cheating whichwe investigate here, we provide a traditionally formatted transcript of theextended interaction between the ITA and his student. The "grade" and"cheating" portions are labeled. This is an impoverished representation of thespoken language data because intonation contours and non-verbal informationare missing. Our purpose here is to provide the reader with the conversationleading up to the "cheating" episode, and also to present our data for the firsttime in a format that the reader may already be familiar with. (For the fullyelaborated transcript of the "discussion of grade" section, along with the analysisof the "inductive/collaborative" strategy, please see Tyler and Davies [1990].)In the version of the transcript presented here we have accommodated spokenlanguage to written language conventions, and we have underlined words thatreceive strong stress so that the reader can have a better idea of how theinteraction sounds. A backslash indicates overlapping speech (as in the T'ssecond turn with "And"), and unclear segments are indicated with squarebrackets (as in the T's second turn after "should").

Discussion of Grade

The student has just asked why he received such a low grade on the homework:

T: Let me see your paper. Well, you've written all these numbers all overyour graph.

S: It says to. It says to. In the other paper it says plot your...it says write thenumbers of your points. That's why I did it.

T: \And let's see...you have no enclosures. You should [ ]...ah..if youhave point here,

S: ah huhT: then you should ...ah...because of this..this..because this is graph...this

[you will] need...connected. And if...with connection., (may, make) .thepoint may not be seen. You should enclose this point with some shape.

S: [Well, I] ... .if you can't see the points there...T: I know, but it is mandatory...for your [graph, factor] analysis. See...see

this one. And graph B.. .maybe this ah.... We have four graphs,so each graph will be four points [ ]

S: \Wai-wai-wait. What about the calculations?T: Oh. In this case I didn't take off any number....any scores...for

calculations, becauseS: \Wai-ai-ai-ai. I wrote three pages of calculations and I'm not going to get

any credit for it?T: No. I mean, I didn't take off.S: You mean you didn't give credit.T: Maybe...maybe.mese calculations is demonstrated on your graph B.

Sorry, you almost missed the important, difficult point of this labassignment. Sorry, but graph B is on log-log paper.

S: It is.T: But you [you didn't] draw this graph on this paper.S: This is log-log paper, isn't it?


T: Is this log-log paper?S: I thought it was.T: Oh, sorry. Hi, hi, Dave, would you please show me the log-log paper? So

there's no point here.(Dave brings paper.)T: Thank you. This one. This and this.S: I thought that was log-log paper,T: Log-log paper means....you know what the log-log paperS: \ I know exactly what it means. I thought that was log-log paper.T: But I can't give you any.. .any numbers.. .any points..for this graph. And if

you want to find slopes, or if you want to ah ana- analyze this graph, youshould ignore this [point].

S: I should. Yeah. I just wrote it in there to show you that I came up_ with it.T: \And but find the slope of this...graph...you should ah...you should

ah..you should have some straight lines. That's the sense of line fitting.Your data points should not be...will not lie on the exact...straight line.They should be scattered [ ] [ ]. But you should.. .you should make.. .a bestguess. Make straight line here. This...look. No sense. It has no sense. Tofind some [][][][][]. The same for this one.

Discussion of Cheating

The student takes out his quiz and asks why the ITA had written "Please do notcheat" at the top. The ITA also has the quiz of the other student who issuspected of cheating. The ITA lays the two quizzes down side by side on thedesk and says something like "Let's look at these quizzes. Let's see what'swritten here."

T: The figures are exactly the same. The symbols you use is exactly thesame.

S: Well you have to use the law of co-sines to get the answerT: \I_knowIknow I know you should butS: \Look at her answers here. They're just like mine. She didn't even have

anything there. I had most of that.T: YBut But I was confident that she or you have cheated.S: haaaahT: Maybe you didn't cheat but ah you are supposed to protect your answers.

Am I right? Know what I mean? You understand? Is there any properreason to this special coincidence?

S: Well, you have to work— everyone worked them out the same, I bet.T: _I know but because of their personalities and their information and

knowledge [ ], there... these two answers should not be (the)same. Butthis is from my experience.

S: So I don't get any credit for the quiz.T: Maybe ah if you just apply for this... if you just apply the quiz, there you

have C. Then you have C. Is equivalent to one point. It's fair.S: I've been reamed.T: Sorry.


TRANSCRIPTION AND ANALYSIS

This section first presents the elaborated transcription of the "cheating"interaction, followed by an analysis that considers several dimensions of thediscourse (nonverbal, prosodic, lexical, syntactic, pragmatic). The discourseanalysis leads to key questions for which a structured process of video playbackand interviews with representative members of the relevant speech communitiesyields crosscultural data concerning norms, schemas, and interpretations of thesituated interaction. This data in turn allows both a fuller understanding of theinteraction and some theoretical conclusions concerning discourse strategies andinterlanguage pragmatics in crosscultural institutional context.

In the following transcription we have tried to represent the speakers'utterances and their nonverbal behavior in an approximation to real time. Eachnumbered horizontal space (containing in reality three lines) represents a stretchof time within which the speakers are interacting. The nonverbal behavior isdescribed on the uppermost line, the ITA's utterances are on the second line, andthe student's utterances are on the third line. We have used standard punctuationto indicate prosody. A falling intonation is signaled by the use of a period (.) anda rising intonation is signaled by the use of a question mark (?). Words orsyllables that receive noticeable stress have been underlined.

There is a short gap in the tape here. Our observational notes and notes frominterviews with the ITA indicate that the ITA responded with something like,"Let's look at these quizzes." The ITA has the quiz of the other student who issuspected of cheating (the missing lab partner), finds it and lays the two quizzesdown side by side on the desk and says something like, "Let's see what's writtenhere." After a few remarks concerning an erroneous written comment made bythe ITA, the discussion of cheating begins:

1.ITA points to two test papers lying side by side on the deskT: The figures are exactly the same. The symbols you use isS:2.

whiny, challenging student has leaned into lowerposition than ITA

T: exactly the sameS: well you have to use the law of co-sines to get the answer3 .

ITA oriented to studentT: J know I know I know you should but butS: Look at her answers here.They're just like mine. She didn't even have anything there I had most of that4.

drop in volume on "cheated"T: but I was confident that she or you have cheated.S: haaaah

Looks away


5.T: Maybe you didn't cheat but ah you are supposed to protect your answers.S:6.lower volume—student does not meet his gaze return to typical volumeT: Am I right? Know what I mean? You understand? Is there any properS:7.

long pause—student does not meet gazeT: reason to this special coincidence?S: Well8.

T: I know butS: you have to work...everyone worked them out the same I bet.9.

T: because of their personalities and their information and knowledge[ ]S:10.

gestures towardown cheststudent avoids gaze

T: There...these two answers should not be (the) same. But this is from myS:11.

ITA orients away from student to grade bookT: experience. Maybe ah if you justS: So I don't get any credit for the quiz.12,

student stands up so that he is taller than ITA, looks at himT: apply for this. If you just apply this quiz, there you have C. Then youS:13.

not addressing ITAT: haveC. Is equivalent to one point. It's fair. Sorry.S: I've been reamed.

Initial Discourse Analysis

In an initial discourse analysis the analyst looks at the video carefully, andpinpoints trouble spots and key questions to be pursued with the participants andother consultants. The primary perspective represented in the discourse analysisat this, point is inevitably that of the analysts, in this case Americans withuniversity teaching experience. For the American analysts, there is a strikingdifference between this interaction and the discussion of grades as analyzed inTyler and Davies (1990). Cheating is bad behavior, and American universitiesare concerned about its apparent increase among students. The suspicion of


cheating charges the atmosphere of the entire conversation with a moral/ethicalweight. The Korean ITA's treatment of the suspicion/accusation of cheating, incontrast to his inductive/collaborative discourse strategy in the discussion ofgrading, is very direct and even confrontational. In dealing with this topic, theITA appears to act in a very authoritarian way, explicitly asserting hisconclusions without allowing the student much opportunity to respond to theaccusation, thus using a very face-threatening version of a deductive/assertivestrategy. Even though the American analysts respond, as does the student, withextreme discomfort to this authoritarian approach, they are also aware that suchresponses to moral/ethical transgressions are in fact part of the repertoire withinAmerican society, and might assume that such responses are more typical inother societies. Such assumptions might lead the analysts to prematurelyconclude that what is happening here is a transfer of discourse strategy from aKorean context to an American one.

It may be useful to have an overview of what we are calling the ITA'sassertive/deductive strategy before we offer the more detailed discourse analysis

• A framing of the interaction by writing the message "Please do not cheat" atthe top of the page on the quiz. It presupposes the fact of cheating on thequiz as well as exhorting appropriate behavior on future quizzes. (There is astrange mismatch of the use of "please" with the imperative concerningcheating. And, in fact, the use of "please" seems incongruous with the ITA'streatment of the cheating situation in his face-to-face interaction with thestudent.)

• A strong statement of the facts that demonstrate cheating (while pointing toquizzes).

• An immediate statement of conviction that one or the other student hascheated.

• A statement that even if this student hasn't cheated, he is still guilty of notshielding his answers.

• Three questions in a row demanding/attempting to solicit agreementconcerning behavior norms in a testing-taking situation.

• One (potentially sarcastic) question demanding(?) an explanation.• A statement of his experience as a teacher to back up his conclusion that

cheating has occurred.• An explicit judgment that his treatment of the situation is "fair."

We suggest that you now follow the transcript closely as you read thediscourse analysis. In lines 1-2 the ITA begins with an observation about thetwo quizzes that constitutes the evidence for his conclusion that the studentshave cheated, and then follows it up immediately with his judgment in line 4. Inhis first two statements concerning the quiz, the ITA uses syntactic parallelismwhich repeats the heavily stressed phrase "exactly the same." There is noequivocation here. When the student responds in line 2 that everyone wouldhave had to use the same "law" to work the problem, the ITA responds in line 2with repetitions of "I know" that sound dismissive to the American ear because


the stress is on "I" rather than "know." The ITA then asserts in line 4 that "I amconfident that you or she have cheated." The ITA does lower his volume on thefinal word, cheated, in line 4. This might be a move to mitigate the face threat tothe student by de-emphasizing the words that explicitly carry the accusation andalso to potentially provide a bit of privacy for the student since the interaction istaking place in front of the other students. Even though the statement might beconsidered to be mitigated in some sense because the possibility is presentedthat the student is not the cheater (but rather that it is the absent female student),the use of the predicate adjective "confident" in line 4 to describe his level ofsureness is quite strong for the ITA.

The unexpected directness of the accusation is registered through thestudent's uptake at that point: He looks away while expelling breath with thesound "haah" at line 5, indicating "seriousness" and strong emotion. It should benoted that even though the ITA states that he is "confident" that one of thestudents has cheated, this utterance is not a direct speech act of "accusation" ofthe student standing in front of him. Neither is there a direct response from thestudent. During this interaction the ITA is physically oriented to the student,who seems to be expressing his ambivalence physically toward the ITA. On theone hand he is expressing deference by leaning over so that his body is literallyat a lower level than the ITA. On the other hand he is expressing disrespect (inAmerican terms) by avoiding eye contact.

The ITA's next utterance in line 5 at first appears to be mitigating ("Maybeyou didn't cheat"), but it is followed immediately by a clause that appears toblame the student for his behavior ("but ah you are supposed to protect youranswers."). The way the ITA frames the situation makes the student guilty inany case (even if he didn't cheat, he didn't shield his answers and therebyallowed the other student to cheat).

Following this utterance, the ITA produces a series of questions in lines 6and 7 ("Am I right? Know what I mean? You understand? Is there any properreason to this special coincidence?") which seem to shift gradually fromrhetorical to authentic questions. The first three questions appear to berhetorically asserting and then seeking confirmation for the claim concerningnormative behavior for test-taking. These three questions are produced at lowervolume with orientation to the student. The final question appears to bedemanding an explanation from the student for the similarity of the answers onthe two quiz papers. Following on the blaming statement about protectinganswers, such a series of questions is highly face-threatening. The lack ofinteractional space granted by the ITA for a response after each of the first threequestions contributes to the intimidating effect. Even though the ITA's finalquestion might appear to presuppose (through the use of the words "specialcoincidence") the student's innocence here, given the previous interaction andstance that the ITA has taken, the final question could sound almost sarcastic toan American ear. The use of the adjective "proper" to modify "reason" is alsopotentially interpretable as sarcastic; it suggests that the student may have"reasons" but not ones judged by the ITA to be proper or satisfactory. Whereas


the utterances appear to challenge the student, there is some ambiguity in theITA's lower volume and physical orientation to the student. It may be the ITA'slack of lexical and grammatical competence that contributes to the challengingquality of the last question.

Discourse-data-internal evidence of the significant impact of this series ofquestions is the length of the pause before the student responds. When he finallydoes respond, it is lamely to assert in lines 7 and 8 that "Well you have to work.. .everyone worked them out the same I bet." The use of the opener "well"suggests a weak disagreement, but the student's shift of subject in mid-clause inline 8 suggests a retreat in response to the strong face-threat presented by theITA. He begins with generic "you," claiming knowledge of how such problemsmust be worked out. He then apparently realizes that he is, in fact, speakinginappropriately to the teacher and backs down from such a claim to authority,shifting to a statement about the students in the class (represented by "everyone"in subject position), which is then further hedged by a final "I bet."

The student avoids responding directly to the four questions the ITA hasjust posed—neither does he respond to the accusation of cheating. The ITAresponds with another "I know" that again sounds dismissive—because of stresson "I" rather than "know"—and then concludes with a final assertion of hisauthority in lines 8-10. He states that "but because of their personalities andexperience and knowledge, there. . .these two answers should not be the same."Even though the initial subordinate clause is unclear (although he appears to begetting at individual differences among students), the final clause is very strong,with the use of the modal "should." This response by the ITA then culminates ina powerful statement at line 10 of his authority as the teacher, even though it isprefaced with "but:" "But this is from my experience." The "but" here soundslike a conventional nod in the direction of self-effacement that is not to be takenseriously.

The ITA asserts his authority through the implication that he has hadextensive experience with students in this sort of situation. This stands incontrast to the way he handles the situation about log-log paper in thediscussion of grades in the previous study (Tyler & Davies, 1990). In the log-log paper situation, when the student challenges his statement that the studenthasn't used log-log paper, he doesn't assert his authority verbally, but ratherasks another student for a sample of log-log paper and then shows the studentthe difference, which the student accepts. A statement, "But this is from myexperience," is unanswerable by the student who clearly cannot claim suchexperience for himself. When the student does respond, he begins his utterancewith "so," linking his statement to the previous discourse by casting thediscussion of the quiz as the premise from which he now draws his conclusion,at lines 10-11, that "I don't get any credit for the quiz." At this point the ITAseems to have put the student in a position of drawing a conclusion andaccepting it. Notice, however, that the conclusion is not "yes, I cheated" or "Ididn't cheat but I clearly failed to shield my answers"—but rather appears to besomething like an implicit acknowledgment of cheating or of failing to shield,


together with an acceptance of the ITA's power in the situation. It is a"conclusion" that acknowledges his defeat in the attempt to get more points onthe quiz. The student initiates a shift here away from the moral to the logistical.

The ITA's next series of utterances are logistical, in terms of points andcourse grade, etc., along with an explicit moral judgment at line 12 concerninghis handling of the situation: "It's fair." The ITA orients away from the studentas he consults the grade book. It is noteworthy that even though the studentdoesn't admit to cheating, the ITA is penalizing the student with points as if hehad cheated. This may be because the ITA has framed it that whether the studentactually cheated or not, he is still guilty of not having shielded his answers.Moving out of his position of physical deference, the student stands up to hisfull height and looks down at the ITA as he hears the grades that he will get. Thestudent's defeat is registered by his loser's comment, not uttered in the directionof the ITA, "I've been reamed," at line 13. This is quite rude, with sexualconnotation, although the student may not be consciously aware of thatdimension of the expression. The passive construction suggests a lack of blameon the part of the student. The student is ultimately penalized in terms of pointsfor the quiz.

Key Questions

At this point we have established from the discourse analysis that the KoreanITA apparently switched his discourse strategy from an inductive/collaborativeone in the discussion of grades to a deductive/assertive strategy in handling thesuspected cheating. He also assumes a much more assertive role whichemphasizes his experience and authority. The next step is to try to discover bothKorean and American norms concerning cheating in order to shed light on theinterlanguage pragmatics dimension of the situation. In the stress of the situationis the Korean transferring a version of a Korean discourse strategy for thisparticular topic and situation? What can American TAs tell us about theirschemas for this situation and what the American undergraduate was probablyexpecting?

Creating Structured Techniques for Answering the Questions. After wehad obtained 1RB approval for the project involving human subjects, we soughtappropriate Koreans and Americans to be interviewed. Potential intervieweeswere told they would be asked about issues of importance to teachers cross-culturally. We used interviews from six Koreans who had some experience ofteaching in university contexts in Korea and who hadn't been too Americanizedby their experience at American universities. We had release forms ready andobtained permission to taperecord the interviews if it seemed appropriate.Otherwise, we took notes on the interviews. Individual interviews wereconducted in university offices, and tape recording was done using a RadioShack tietack microphone positioned between the interlocutors. For theAmerican TA interviews we selected one university English teacher with 20


years of experience, one with about 5 years of experience, and one who had justbegun as TA the previous semester and thus had just recently receivedorientation as an American university teacher. The interview protocols areincluded in the appendix. We created questions that moved from abstract (inorder to get at norms) to very specific and situated when we asked theinterviewees to look at the videotape and comment on the Korean ITA'shandling of the suspected cheating. The questions were intended as generalguides and prompts to initiate open-ended responses. Challenges were toprovide appropriate contextualization for the transcript and videotape, and toprepare the interviewee for viewing the tape. We read descriptions of the contextand immediately preceding situation to the interviewees, in order to createconsistency across interviews. Anticipating that the transcript as given abovemight be confusing in the interview situation, we created a less detailed versionof the transcript, treating it as a screenplay. While the interviewees followedalong with the transcript, we read it aloud to them, maintaining stress patterns asmarked with underlining and generally imitating the ITA and student. We askedfor any questions on points needing clarification, and then we played thevideotape. Interviewees were asked to comment on anything that struck them.Comments were followed up in relation to specific moments in the interaction;for example, a comment that "the ITA doesn't give the student a chance to talk"would be referred immediately to the transcript to find what was happening inthe discourse from the interviewee's point of view and what aspects of thediscourse were being used as contextualization cues. For example, theAmericans confirmed our analysis that the stress on "I" rather than "know,"together with the lexical choice of "know" rather than "understand" on the partof the ITA were significant in creating the impression that he was dismissing thestudent's response rather than expressing sympathy at line 2.

Findings From the Representatives of the Speech Communities

The interviews served both to validate and further inform our initial discourseanalysis and also to provide rich data on the relation between discourse strategyand topic in the institutional context of the university. Interestingly, both theKorean and the American teachers commented that a more skillful teacherwould have prevented this situation from ever arising, and discussed appropriateclassroom management techniques.

Americans. For the Americans, the data represent a situation fraught withmoral and ethical issues; "cheating" is considered to be morally wrong. There isconcern that university students are engaging in more and more of this behavior,and American universities are taking up "academic misconduct" as an importantissue. The Americans expressed shock at the "confrontational" way that theKorean ITA handled the discussion of potential cheating. All of them said thatthey would handle a comparable situation in a very indirect way, with someversion of what we have called an "inductive/collaborative" discourse strategy.


It was clear from their responses that the Americans were affected byideological, cultural, and institutional constraints. The ideological dimension (interms of the basic principle of "innocent until proven guilty") was least overt butappeared to motivate a reluctance to make a direct accusation in thecircumstances. The Americans also expressed a sense of the student's privacybeing violated and noted that such a discussion would never happen in front ofthe entire class. A related issue was the cultural fact of the current litigiousnessof American society; teachers are now worried about being sued over statementsthat they make to students. The institutional constraints are most obvious in thecurrent trend (in place at both of the researchers' universities) to remove theteacher from any confrontation with the student in cases of suspected academicmisconduct. The least-experienced teacher, who had just gone through TAorientation, seemed most aware of current policies concerning how she shouldhandle cases of suspected academic misconduct. A teacher is explicitly directedto avoid any situation like the one in the data, submitting evidence of cheating toa departmental or college committee. Policies governing the handling ofacademic misconduct through departmental and university review boards can belocated on American university websites.

Thus we see that the constraints imposed within the culture of an institution(here as part of a national trend) can affect discourse strategy in a profound wayin relation to a particular topic. Whereas we might assume a cultural preferencefor the "deductive/assertive" style in many situations, given overt Americannorms for speeches and expository essays, we need to be careful aboutinappropriate generalization. When the topic is cheating, Americans appear toopt for an "inductive/collaborative" style. Even when one of the Americansresponded initially that she would "confront" the student privately with theevidence of cheating, probing revealed that what she really meant was that shewould simply show the evidence to the students and ask them to explain, butwould offer no verbal conclusions (i.e., she would opt for an extremely indirectform of the inductive/collaborative strategy).

Koreans. The Koreans consistently emphasized that cheating is treatedvery seriously, especially in secondary school because the competition to getinto good universities is stiff. They also noted that, unless the student uses hightech, cheating is difficult. High school teachers go to elaborate lengths toprevent the possibility of cheating. Students are often physically positioned sothat they are sitting at a maximal distance from one another. Two of ourinterviewees related experiences in which teachers had students stack backpackson chairs situated between students in order to create physical barriers betweenstudents as a way of curtailing the possibility of students looking at each other'spapers. In addition, they reported there were often additional proctors roamingthe classroom, looking for any signs of cheating. Three of the informants saidthat teachers also give specific instructions on how individual students shouldguard their work so that other students cannot see their answers. Two noted they


had seen teachers take away test papers if students were sitting in such a waythat it appeared they were allowing other students to see their answers.

In response to the question of what Korean teachers do when they haveevidence of cheating but have not actually seen the students engaged incheating, the Korean participants unanimously said the teacher would mostlikely treat the situation cautiously. One informant said he felt that the strengthof the teacher's reaction would be in proportion to the importance of the test orassignment. He noted that small quizzes are unusual in Korea, but if theassignment were not worth much, the teacher might not react strongly tosuspected cheating. The Koreans reported a range of responses they had heard ofor observed. The teacher might caution the entire class about cheating and thenincrease the vigilance in looking for cheating by bringing in extra proctors ordividing the class into smaller groups and testing each group in differentclassrooms, but not specifically accuse the suspected parties. As a way to createa group reaction against cheating, the next quiz would often be much moredifficult than the previous one. They uniformly reported that they had neverheard of a Korean teacher writing a message like "Do not cheat" at the top of atest. Neither had they ever heard of a teacher publicly accusing a student ofcheating if the teacher had not actually observed the student in the act ofcheating.

The Korean interviewees also stressed that a Korean teacher had much moreauthority than an American instructor and if a teacher determined that studentshad cheated, he or she might simply give the student an F or a lowered gradewith no discussion. In many instances quizzes and tests are not returned, so thestudents are not always sure as to why they receive a particular grade. Thepresumption is that if the students had cheated, they would accept the lowergrade without question. If a student thought he or she had done well andreceived a lower grade than expected, an option might be to privately andhumbly ask the instructor about the grade. But in this situation, the studentwould not challenge the teacher's assessment, so the question of cheating mightnever arise.

Another practice is that the teacher might tell the suspected students to"follow him to his office." The interviewees noted that this command in itself isoften considered a serious reprimand. Once in the office, the teacher wouldlikely show the students the written material that made the teacher suspectcheating. In this case, the strong expectation would be that the student(s) wouldimmediately confess without the teacher having to explicitly accuse the student.In other words, if the instructor chose to discuss cheating, the preferreddiscourse management strategy would be an inductive, nonconfrontational one.In some ways the strategy is not significantly different from the typicaldiscourse strategy employed by American instructors. There are some importantdifferences, however, primarily having to do with the authority of the teacher tomake a determination of cheating and act on it without informing the student.

As we reported in Tyler and Davies (1990), the ITA felt confused abouthow to handle the entire interaction. His initial instinct was to postpone the


discussion of grades until after the class. However, he had also been warned byother Koreans that American students ask many questions in the classroom(something that is unusual in Korea) and that the teacher is expected to answerthe questions. He was unsure if it would be considered appropriate to postponethe discussion until after class and so determined that he should answer thestudent's questions during the lab, when the other students were present.

Although the ITA had attempted to make adjustments to the informality ofthe U.S. classroom (he dressed in shorts and a tee shirt and joked with thestudents) and to the greater consideration/authority given American students (headdresses students by their first names and elicits their questions), in the videoplayback sessions he stated that he was taken by surprise at the student'srepeated rejections of his explanations for the student's poor grade on thehomework assignment. He was not prepared for the student to so boldlyquestion the instructor's judgment and authority. He clearly articulated a senseof growing tension during the interaction, pointing out places on the videowhere he held his back more and more stiffly as the exchange continued.Although he attempted to pursue the goal of calmly explaining in detail theproblems with the student's homework, after 10 minutes he felt that the studentwas being unduly quarrelsome.

It was at this point that the issue of cheating arose. Based on his experiencein Korea, he never thought a student would question a teacher's judgment aboutcheating. Thus, he found himself in a situation he would never encounter inKorea.

In our initial interviews, we asked the ITA why he had written "Please donot cheat" on the quizzes. He stated that the quiz was a small part of the overallgrade and while he wanted to discourage further cheating, he did not feel thatcheating on the quiz warranted a discussion in his office. This squares with ourother informants' remarks about the seriousness associated with being told to goto the teacher's office. This ITA wanted to warn the students not to cheatwithout penalizing them too severely. He pointed out that, rather than giving thestudents zeros, he gave some credit for the quiz. He was hesitant to alert theentire class to the fact that he suspected some of the students of cheating, as hewas aware that American teachers were expected to build friendly relations withstudents and such a general warning might impair that rapport. His sense wasthat a written warning given individually would be a face saving move thatwould avoid involving the entire class.

DISCUSSION

Our triangulated data suggests that the discourse management strategies used bythe ITA in this interaction cannot be explained simply as a matter of negativetransfer from the L1/C1. In fact, we haven't found evidence that a Koreanteacher, even with greater authority than an American teacher, would use thisoverall assertive/deductive strategy in a comparable situation. We hypothesizethat the ITA found himself in both a physical and cultural setting that was


substantially different from what he had encountered in Korea. Moreover, thestudent-teacher interaction in which he found himself was quite different fromthose he had encountered in his native culture. There is evidence of attempts onhis part to adjust his discourse management strategies and his enactment of hisparticipant role as teacher to the norms of the U.S. classroom, but it is also clearthat his understanding of those norms was incomplete. Although he drew oncertain aspects of his L1/C1 schemas, it seems that many aspects of hisdiscourse management strategies are outside the norms of either his L1/C1 or thetarget U.S. language/culture. Thus, we see several instances of true interlan-guage pragmatics.

One of the clearest instances and arguably the most important in terms ofthe failure of the interaction, is his decision to respond to the indirect, writtenevidence of cheating by writing "Please do not cheat" at the top of the students'papers, issue partial credit for the quiz, and return the papers without planning todiscuss the issue with the accused students. The message presupposes the fact ofcheating on the quiz as well as exhorting appropriate behavior on future quizzes.The enactment of this strategy serves as a framing for the ensuing interactionwith accused student.

The circumstances that led the ITA to use this strategy are complex. Ingrading the quizzes, the ITA found himself in a situation that would rarely occurin Korean classrooms because of the difference in physical arrangements whenquizzes and tests are administered. Our Korean informants uniformly told us thatvarious physical precautions would be taken to lessen the possibility of a studentcopying from another. The strategy of physically separating the students is notavailable to the ITA in the U.S. physics lab. Moreover, in the Korean context, ifstudents are suspected of cheating, the teacher is likely to bring in additionalproctors to carefully monitor the students and keep further cheating fromoccurring. Again, this is not a possibility in the U.S. situation.

The Korean informants also told us that the teacher might issue a warning tothe entire class if the cheating were considered serious. After such a warning thenext quizzes would be more difficult. However, the ITA was reluctant to issuesuch a warning because of the potential negative effect on his rapport with theentire class, especially since only two students out of the entire class wereinvolved. Moreover, small, regular quizzes seem not to be typical in Koreanclasses, while they are an institutional requirement in the course the ITA isteaching. In the Korean situation, tests that carry more weight seem to be thenorm. Thus, the ITA is again in a situation in which there is no analog in hisnative culture. One of our informants suggested that how strongly the teacherreacts to an instance of cheating is somewhat dependent on the importance ofthe test itself. Thus, another factor in the decision not to talk to the studentsabout the cheating seems to be that the weight of the quiz was so small in theoverall grade. Nevertheless, the ITA does not want the cheating to continue. TheITA seems to have made the decision that the quiz grade was not significantenough to warrant a public warning to the entire class and opts for the privatewritten warning. All the Korean informants indicated that this was a strategy


that was outside the norms of the Korean classroom. Thus, we see a strategyemerging which is influenced by certain aspects of Korean academic culture butwhich plays out in a discourse strategy that would not typically be found in theKorean classroom.

The norms for the participant roles of teacher and student in the Koreancontext are much different from those of the U.S. context. In this area, the ITAseems to have relied on his understanding of the teacher having absoluteauthority in deciding the students' grades and in making the determination thatcheating has taken place. Recall that our informants told us that a Koreanteacher might give a student a low grade because the teacher had determinedcheating had taken place. In such cases, many Korean teachers might opt not todiscuss the issue with the suspected student. If the test were not returned, whichis apparently often the case, the student might not ever be aware that the teacherhad made that determination. The ultimate strategy followed by the ITA in thisinteraction was to make the determination that cheating had occurred (which hefelt was unquestionably supported by the written evidence) and act on hisauthority by issuing a lower grade to both students. In making thisdetermination, he did not feel compelled to find out which student had actuallydone the copying. Here his thinking seems to be largely determined by thestrong Korean ethic that each student is responsible for shielding his or heranswers and that students who do not do so are judged as complicitious. Theseelements of the strategy seem to involve transfer from the norms of the Koreanclassroom. By giving a written warning, he is indicating to the students that he isaware that some cheating has occurred and that he wants it to stop. However, incontrast to the Korean situation, the quiz had to be returned to the students. Inreturning the quizzes with the written warning, the students are made aware ofthe teacher's assessment. This creates a situation different from either the typicalKorean or U.S. norm.

For the U.S. student, any allegation of cheating constitutes a serious threatto the individual's personal integrity. In this situation, even a guilty student mayfeel the need to defend his or her honor. The Korean informants suggested that ifthe student were innocent, he or she might privately discuss the matter with theteacher, but if the teacher were firm in his or her decision, the student wouldacquiesce. The guilty student would likely stay quiet and accept the lower grade.Thus, the ITA is taken by surprise when the student publicly asks for anexplanation. (In truth, we find the student's behavior to be outside the typicalnorms of the U.S. classroom. Our sense, which was confirmed by interviewswith the U.S. informants, is that this type of conversation is most likely to takeplace in private, at least in part because the student wants to avoid beingembarrassed in front of the other students.) In the Korean context, the studentwould never ask the question, especially after having a testy discussion of ahomework assignment. Once again, the ITA is in unknown waters. Stillfollowing his filtered understanding that in the U.S. classroom the teacher has toanswer the student's question, he forges ahead with a strong statement of thefacts that demonstrate cheating (while pointing to quizzes), "The figures are


exactly the same. The symbols you use is exactly the same." The accusation ofcheating followed by explicit presentation of the evidence on the part of theteacher represents a deductive discourse management strategy which strikes theU.S. listener as confrontational. It also struck the Koreans as unusual. Theysuggested that the teacher would show the two quizzes to the two students andwait for one or both to confess.

At this point, the student challenges the teacher's evidence, "well, you haveto use the law of cosines to get the answer." The Korean informants indicatedthat students in Korea would never offer such a challenge. Even in the face ofsuch a confrontation, the ITA still attempts to be responsive by indicating that heunderstands the student's point of view. His lexical choice, however, of "know"rather than "understand," and a stress on "I" rather than on the verb, creates animpression that he is being dismissive rather than sympathetic. Being in theunexpected position of having his judgment and authority directly challenged,the ITA, becomes even more direct, "But I was confident that she or you havecheated." This statement is followed by a rapid fire series of questions that havethe effect of aggressively challenging the student to offer an explanation. OurKorean informants were very uncomfortable with this exchange. They saw theITA as aggressive, but they also saw the student as disrespectful. The ITAhimself revealed that he was surprised and somewhat insulted by the student'schallenge and felt the need to defend himself. Again, it seems that the ITA isimplementing discourse strategies that only partially reflect his L1 pragmatics.The strategy of aggressively pointing out the evidence of cheating and demand-ing that the student offer an explanation is outside the norms of Korean teacherbehavior largely because the dynamics of the Korean classroom are such that thestudent would never offer such a challenge. Simultaneously, the ITA isinfluenced by his sense that the teacher should not be challenged and that as theteacher he had the right to make the assessment that cheating had occurred. Oneof our Korean informants explained that some "scary male teachers" in highschool are very aggressive and even use corporal punishment if they feelstudents are disrespectful. Thus, the discourse management strategies thatemerge, while reflecting aspects of the Korean schema of participant roles, alsodiffer from typical Korean classroom behavior.

An irony of the situation is that "positive transfer" of a collaborative/induc-tive discourse strategy by the Korean ITA might have provided a better matchwith typical American behavior in this context.

Existing Institutional Processes as a Source of Data

The Use of Videotaping in the Preparation of Teachers. With theprevalent use of videotaping as a tool in the preparation and supervision ofteachers, such tapes can be an important source of institutional talk as data forresearch on interlanguage pragmatics. Bardovi-Harlig and Hartford (chapter 1,this volume) provide information on the logistics of this important tool. Perhapsthe most complex use of videotaping in teacher preparation and supervision is


the kind of process out of which this particular data emerged. In this model,routine taping and playback methodology (cf. Davies & Tyler, 1994; Davies etal., 1989; Tyler & Davies, 1990) are an integral part of the process.

The Kinds of Data Created. Videotaping and analysis of the sort describedwould provide data for at least three sorts of studies within the framework ofinterlanguage pragmatics (cf. Bardovi-Harlig, 1999): (1) longitudinal data withindividual IT As; (2) data to track the effects of specific instruction; (3) anopportunity to match taped data of actual teaching performance with other sortsof evaluative data, in particular with student evaluations of teaching. Theexamination of actual teaching performance in such cases might yield someinteresting findings concerning the relative importance of pragmatic competencein relation to other sorts of competence (i.e. that an ITA with well-developedpragmatic competence might be able to compensate very effectively for a lack ofcompetence in other areas of language—pronunciation, grammar, lexicon, etc.,as we saw in the case of Wes [Schmidt, 1983]). And that an ITA with excellentpronunciation, for example, but lack of pragmatic competence might get verypoor teaching evaluations from students. This would challenge "commonsense"notions about language proficiency and interactional competence.

Existing Institutional Processes as Part of a Methodology

Triangulation of Multiple Perspectives. Institutional arrangements forteacher supervision that involve discussion of the videotapes also provide animportant source of data for methodologies that require triangulation from theperspectives of the participants in addition to the perspective of the analyst.Playback methodology with the videotapes also allows the elicitation of inter-pretations from other members of the interpretative communities represented bythe participants. This is key to providing crucial insights that the analyst, limitedby his or her own cultural assumptions, does not have access to from thetranscript alone. Such a methodological procedure also helps to tease out thecultural from the individual.

The Identification of Key Problematic Situations. Routine collection ofinstitutional talk (i.e. systematic gathering of natural data in a range ofinstitutional contexts) allows the researcher to identify key problematicsituations where communication breaks down. The value of such data from thistheoretical perspective is that it allows the analyst to use miscommunication forinsight into the taken-for-granted schemas underlying successful communication(Erickson & Shultz, 1982; Gumperz & Tannen, 1979). This orientation to thedata contrasts with a "strategic" approach (Canale & Swain, 1980) in which thefocus is on discovering what learners do to compensate for miscommunicationwith the assumption that their linguistic behavior in these contexts is importantto the process of language acquisition. Our assumption, in contrast, is that thelearner is unlikely to "acquire" more targetlike pragmatic competence or gain


insight into the state of his or her interlanguage pragmatic competence withoutinstruction of a sort that is much more like a process of explicit socialization (cf.Bardovi-Harlig et al., 1991; Davies, in press; Davies & Tyler, 1994; Gumperz &Roberts, 1980; Tyler, 1994). Analysis of data of miscommunication in keyproblematic situations allows insight into both crosscultural and interlanguagepragmatics, and such insights can be used in preparing curriculum for instructionof the sort just described (cf. Tyler, 1994). In this sort of instruction thevideotapes are a central element, subject to joint analysis by the supervisor andteacher/learner.

CONCLUSION

In analyzing this example of interlanguage pragmatics and explicating ourmethodology, we hope to have addressed limitations of current research ininterlanguage pragmatics in terms of methodology, data, and theoreticalframework. Through our interactional sociolinguistic analysis of this discoursedata we have shown that simple notions of "transfer" are not adequate tounderstanding interlanguage pragmatics. Rather, each crosscultural situation ispotentially a new context, a "third place" (Kramsch, 1993) in the interfacebetween cultural and linguistic systems. We have seen that the L2 learner in ourdata produced discourse that was a complex construction built of L1 pragmaticsand of a partial understanding of the target language pragmatics filtered throughhis perception of the target culture. The discourse was further shaped byresources and constraints of particular institutional contexts.

We also hope to have established the notion of "discourse strategy" as adimension of interlanguage pragmatics and as an interesting area for research.We take "strategy" to mean a purposeful linguistic maneuver that provides bothan overall ordering to the shape of an extended stretch of discourse, as well asmore local stratagems for accomplishing a particular communicative goal inspoken interaction. Each language/culture privileges particular discoursestrategies in particular situations. Like other aspects of language, L1 discoursestrategies can potentially be transferred in L2 production (as demonstrated inTyler & Davies, 1990). Understanding the circumstances under which suchtransfer takes place, with either felicitous or infelicitous results, is an importantcomponent of bettering our understanding of the development of pragmaticcompetence. Conversely, understanding the circumstances under which an L2learner uses discourse strategies which match neither the L1 nor the L2 addsimportantly to our understanding of ILP.

Within an analytical approach to interlanguage pragmatics/discourseanalysis it might seem "logical" to separate out different levels of language ordimensions of competence and then try to study them separately. Such ananalytical approach is presumably based on the assumption that, for example,the learner builds up from grammatical to pragmatic competence, or that thereare certain a priori elements that can be isolated and studied (e.g., speech acts),which are then strung together to create a "discourse strategy." In contrast, we


hope to have demonstrated that a holistic approach to the data is essential, onethat makes clear the potential role of all levels of linguistic organization in"interlanguage pragmatics" and the importance of context at the cultural andinstitutional level as part of the interpretive process.

Finally, we also hope to have demonstrated that naturally occurringinstitutional talk can be a valuable and accessible resource for researchers.

References

Bardovi-Harlig, K. (1999). Exploring the interlanguage of interlanguage pragmatics: A researchagenda for acquisitional pragmatics. Language Learning, 49 (4), 677-713.

Bardovi-Harlig, K., Hartford, B. A., Mahan-Taylor, R., Morgan, M. J., & Reynolds, D. W. (1991).Developing pragmatic awareness: Closing the conversation. ELT Journal, 45, 4-15.

Bialystock, E. (1983). Some factors in the selection and implementation of communication strategies.In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication (pp. 100-118).New York: Longman.

Bruner, J. S. (1981). The social context of language acquisition. Language and Communication, 1,155-178.

Canale, M., & Swain, M. (1980). Theoretical bases for communicative approaches to secondlanguage teaching and testing. Applied Linguistics, 1, 1-47.

Connor, U. (1996). Contrastive rhetoric: Cross-cultural aspects of second-language writing. NewYork: Cambridge University Press.

Corder, S. P. (1983). Strategies of communication. In C. Faerch & G. Kasper (Eds.), Strategies ininterlanguage communication, 15-19, New York: Longman.

Davies, C. E. (in press). Developing awareness of crosscultural pragmatics: The case of American/German sociable interaction. Multilingua: Journal of Cross-Cultural and InterlanguageCommunication.

Davies, C. E., & Tyler, A. (1994). Demystifying cross-cultural (mis)communication: Improvingperformance through balanced feedback in a situated context. In C. G. Madden & L. M.Cynthia (Eds.), Discourse and performance of international teaching assistants (pp. 201-220).Alexandria, VA: TESOL Publications.

Davies, C. E., Tyler, A., & Koran, J. J. (1989). Face-to-face with English speakers: An advancedtraining class for international teaching assistants. English for Specific Purposes, 8, 139-153.

Dulay, H. C., & Burt, M. K. (1975). A new approach to discovering universal strategies of childsecond language acquisition. In D. P. Dato (Ed.), Developmental psycholinguistics: Theory andapplications (pp. 209-233). Washington, D.C.: Georgetown University Press.

Duranti, A., & Goodwin, C. (Eds.). (1992). Rethinking context: Language as an interactivephenomenon. New York: Cambridge University Press.

Ellis, R. (1994). The study of second language acquisition. New York: Oxford University Press.Ellis, R. (1999). Learning a second language through interaction: Studies in bilingualism, Vol 17.

Philadelphia: Benjamins.Erickson, F., & Shultz, J. (1982). Counselor as gatekeeper: Social interactions in interviews. New

York: Academic Press.Faerch, C., & Kasper, G. (Eds.). (1983). Strategies in interlanguage communication. New York:

Longman.Green, G. M. (1996). Pragmatics and natural language understanding (2nd ed.). Mahwah, NJ:

Lawrence Erlbaum Associates.Grice, H., P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and

semantics: Speech acts, Vol. 3 (pp. 41-58). New York: Academic Press.Gumperz, J. J. (1982). Discourse strategies. New York: Cambridge University Press.Gumperz, J. J. (1992). Contextualization cues and understanding. In A. Duranti & C. Goodwin

(Eds.), Rethinking context (pp. 229-252). New York: Cambridge University Press.


Gumperz, J. J., & Tannen, D. (1979). Individual and social differences in language use. In W. Wang& C. Fillmore (Eds.), Individual differences in language ability and language behavior (pp.305-325). New York: Academic Press.

Gumperz, J. J., & Roberts, C. (1980). Developing awareness skills for inter-ethnic communication.(Occasional Paper No. 12). Singapore: SEAMEO Regional Language Centre.

Hall, J. K. (1995). 'Aw, man, where you goin'?': Classroom interaction and the development of L2interactional competence. Issues in Applied Linguistics, 6(2), 37-62..

Haastrap, K., & Phillipson, R. (1983). Achievement strategies in learner/native speaker interaction.In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication (pp. 140-158).New York: Longman.

Hinds, J. (1987). Reader versus writer responsibility: A new typology. In U. Connor & R. Kaplan(Eds.), Writing across languages (pp. 141-152). Reading, MA: Addison-Wesley.

Hymes, D. (1974). Foundations in sociolinguistics: An ethnographic approach. Philadelphia: TheUniversity of Pennsylvania Press.

Kaplan, R. (1966). Cultural thought patterns in intercultural education. Language Learning, 16, 1-20.

Kasper, G. (1992). Pragmatic transfer. Second Language Research, 8, 203-231.Kasper, G., & Schmidt, R. (1996). Developmental issues in interlanguage pragmatics. Studies in

Second Language Acquisition, 18, 149-169.Kramsch, C. J. (1993). Context and culture in language teaching. Oxford, England: Oxford

University Press.Kramsch, C. J. (1986). From language proficiency to interactional competence. The Modern

Language Journal, 70(4), 366-372.Markee, N. (2000). Conversation analysis. Mahwah, NJ: Lawrence Erlbaum Associates.Oxford, R. (1990). Language learning strategies: What every teacher should know. Rowley, MA:

Newbury House.Saville-Troike, M. (1989). The ethnography of communication: An introduction (2nd ed.). Malden,

MA: Blackwell.Selinker, L. (1972). Interlanguage. IRAL, 10, 209-231.Schmidt, R. (1983). Interaction, acculturation and the acquisition of communication competence. In

N. Wolfson & E. Judd (Eds.), Sociolinguistics and second language acquisition. Rowley, MA:Newbury House.

Trosberg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11, 147-167.Tyler, A. E. (1994). Effective role-play situations and focused feedback: A case for pragmatic

analysis in the classroom. In C. G. Madden & C. L. Myers (Eds.), Discourse and performanceof international teaching assistants (pp. 116-133). Arlington, VA: TESOL Publications.

Tyler, A. E., & Davies, C. E. (1990). Cross-linguistic communication missteps. Text, 10, 385-411.

6

English for Specific Purposesand Interlanguage Pragmatics

Elaine TaroneUniversity of Minnesota

Although research on interlanguage pragmatics (ILP) has made substantialstrides over the last decades, it is my thesis in this chapter that ILP research canbenefit by incorporating central constructs of English for Specific Purposes(ESP), specifically discourse community and genre, and by awareness ofresearch findings in ESP.

SIMILARITIES BETWEEN ILP AND ESP RESEARCHAPPROACHES

Pragmatics examines the way in which speakers and hearers communicate morethan is explicitly said. This ability to infer meaning is particularly strong whenspeaker and hearer share experience as members of the same social group orspeech community (Yule, 1996). Research on interlanguage pragmaticsexamines the way in which members of a native language speech communityacquire the pragmatic speaking norms of a target language speech community.

Research on English for specific purposes examines ways in whichmembers of particular discourse communities use language varieties (genres) tocommunicate with one another in their pursuit of common professional or work-related goals. Pragmatic effects are often at the center of ESP research, whichstresses the way discourse communities agree to use language for work-relatedpurposes. As Widdowson (1998) pointed out, the study of English for specificpurposes is inherently a study in pragmatics. The reason why insiders canunderstand "special purpose genres" not intelligible to outsiders is that they are"able to infer the relevant discourse because of their professional competence asmembers of this discourse community." The discourse is shaped the way it isbecause members of the discourse community, like all speakers, "designutterances to key into the context of recipient knowledge in the most economicalway" (p. 4). Thus, special purpose genres have their origins in pragmaticprinciples of communication, and it is because of pragmatic principles that theyare processible on an ongoing basis by members of the discourse community.

157

158 Tarone

DIFFERENCES BETWEEN RESEARCH APPROACHESIN ILP AND ESP

Intel-language Pragmatics Research

In research on interlanguage pragmatics, the subjects are normative speakers(NNS) of the target language, and are usually compared with native speakers(NS) of the target language with regard to their pragmatic knowledge andperformance. Learners are viewed as moving from one speech community withnative language norms to a new speech community with target language norms.

An idealized construct of "speech community" has been central in the studyof ILP. Although definitions of the construct "speech community" vary slightly,here we can follow Hymes (1972): A speech community is a group of peoplewho share conventions of speaking and interpretation of speech performance.Within the English-speaking world, there are of course many speechcommunities. Preston (1989) argued, for instance, that second-languageclassrooms can be viewed as speech communities, with internally agreed-uponformal and informal speech styles appropriate for use with teachers vs. students.However, in the interest of efficiency, in ILP research it seems to have beengenerally agreed to treat as a single speech community, with a single set ofpragmatic norms, all speakers of one of the standard varieties of English (e.g.,American English). Although Beebe and colleagues (e.g., Beebe & Takahashi,1989) have sometimes dealt with speech acts like rudeness in the context of asmaller speech community such as New York, participants' membership indiscourse communities or more specialized speech communities within theidealized target language speech community has not generally been a factor inILP research. We can exemplify this stance with the following statement: Arequest is a request, across all social contexts; NNSs must master the generalnorms of NSs of the idealized speech community.

A perennial issue for discussion by researchers in interlanguage pragmaticshas to do with the sort of data one should use in this sort of study. If one isstudying a particular speech act, should they limit their data collection to thetaping of natural oral interaction? Such data collection methods, used forexample by Bardovi-Harlig and Hartford (1996) in examining the pragmaticfeatures of academic advising sessions, have the virtue of face validity: Theyshow how learners actually behave in the real world. In addition, they allowresearchers to go beyond documentation of general group patterns, as theypermit the researcher to analyze deliberate individual choices to flout pragmaticnorms (e.g., Broner & Tarone, 2001; Nelms, 2002; Rundquist, 1990; Tarone,2000). However, such studies may also have certain disadvantages:

• they may be extremely time-consuming and inefficient, since the targetspeech act or pragmatic behavior may not occur very frequently in generalsocial interaction;

ESP and ILP 159

• there is a problem of comparability, of the learners with each other and of thelearners with speakers of the target language variety;

• there is a problem of generalizability, since only a few learners can beobserved at a time and it is unclear whether these are representative; and

• there is a competence/performance problem, since observations alone recordbehavior, but not the learner's or the fluent speaker's perspective on behavior.

As a result, some researchers have relied on the use of tools such asdiscourse completion tasks (DCTs) to access a larger number of learners' andnative speakers' introspections about the speech acts which ought to beperformed in a range of hypothetical speaking situations. Such techniquesobviously have their own opposite shortcomings:

• They use taxonomies and frameworks structured by the researcher that maybias the results in unforeseen ways;

• they may permit subjects too much time for reflection and for self-flattery,and thus not predict actual behavior;

• they provide insufficient contextualization so that study participants are askedto respond with idealized responses to idealized situations;

• they emphasize group norms, with no ability to analyze individual choices,such as sarcasm (e.g., Nelms, 2002) and other forms of language play (e.g.,Broner & Tarone, 2001; Tarone, 2000).

In the study of pragmatics, social context is critical, and it is imperative tofind ways to study pragmatics in natural interactions, using techniques thatminimize the shortcomings outlined above.

In sum, work in ILP focuses on participants' status as native speakers ornormative speakers, on "general pragmatic norms" of an idealized TL speechcommunity (as discussed on p. 157 of this chapter), and on the degree to whichindividual second language learners conform to those general pragmatic norms.ILP research has tended to rely on data elicitation devices and not observation ofnatural interaction in social contexts.

English for Specific Purposes Research

Research on English for Specific Purposes, in contrast to what has beendescribed earlier, focuses on the norms of strictly delimited discoursecommunities, the grammar and pragmatic characteristics of genres used withinthose communities, and on the distinction between expert and noviceperformance.

Research on ESP is quite diverse, and includes both corpus linguistics andgenre analysis. In this chapter, I focus on genre analysis as this approach hasevolved in the field of English for Specific Purposes, particularly following themodel of John Swales. LSP (Languages for Specific Purposes) and ESP (Englishfor Specific Purposes) are fields of study focusing on the description of thelanguage production and judgments, not of second language learners, but of

160 Tarone

expert speakers of some language variety or genre used by members of aspecific discourse community. The term discourse community is used indifferent ways in different fields of study. In this chapter, discourse communityis defined, following Swales, as a specialized type of speech community definedsolely by shared common public goals—often professional or vocational goals.Importantly, the discourse community is defined, not as a group of people whoseshared linguistic conventions derive from shared ethnic background, orgeographical space, or social class. The discourse community is defined morenarrowly by the common goals the members of the community share. Typicalexamples of discourse communities are professional groups, academicdepartments, and hobbyists. Swales (1990) gives the example of his stamp-collecting club, which consists of people who are brought together by theircommon goal of collecting stamps. In pursuing that common goal, members ofa discourse community use genres to communicate with one another infurthering their common purpose; in Swales' example, there is a club newslettergenre in which discourse community members advertise various stamps theywish to sell or buy, using technical terms and phrases whose meaning is opaqueto outsiders but crystal clear to them.

The essential linguistic assumption underlying the notion of genre is thatwhen a discourse community uses language for a specific purpose such aswriting a research paper in microbiology, a technical manual for a computer,nursing notes, or a stamp-collectors' newsletter—or for oral activities such as ahistory lecture—that language becomes specialized in both information structureand linguistic form. The discourse community agrees that information should beorganized in a particular way in that genre, and that linguistic forms should beused in mutually agreed-upon ways to signal that organization. As Hymes wouldsay, members of the same discourse community share the same conventions forwriting or speaking the genres that are used by that community in pursuit of itscommon public goal.

The focus of research on LSP or ESP is empirical and practical: It describesthe characteristics of these different genres, or language varieties, so as toestablish what it is that novices to the discourse community need to learn. LouisTrimble often pointed out that no one is a native speaker of any variety ofEnglish for specific purposes; everyone has to learn new norms when joining anew discourse community as an adult (personal communication, 1971). And,increasingly, expert members of an English-speaking discourse community mayor may not be native speakers of English. The important distinction for ESP ison the expert-novice distinction, not the NS-NNS distinction, and on what it isthat novices need to learn about the forms used in target genres to performpragmatic functions. We can exemplify the general stance in ESP research withthe following statement: A request may fail or succeed in a discourse communitydepending on whether its realization fits genre norms. All novices (nativespeakers or not) must master genre norms.

ESP and ILP 161

Let us now turn to an exploration of the twin constructs of discoursecommunity and genre and their relevance for research in interlanguagepragmatics.

THE CONSTRUCTS GENRE AND DISCOURSE COMMUNITY

The construct of genre was memorably defined by John Swales (1990, p. 58) asa class of communicative events used by a discourse community to serve someagreed-upon set of communicative purposes. Those purposes have an impact onthe information structure and the linguistic characteristics of the genre.

The discourse community itself is a group of people with shared commonpublic goals. The discourse community has expert members who know thecommunity norms and can transmit them to novices. Expert members of thediscourse community share common assumptions about the purposes andstructure of genres used by that community. Members of the discoursecommunity use genres as a means of communication to pursue their commongoals.

Use of these twin constructs, discourse community and genre, can enhanceILP research in several ways. First, they can make it possible for researchers oninterlanguage pragmatics to gather natural data in highly comparable socialcircumstances, and thus to compare the performance and perception of expertsand novices, native speakers and normative speakers. The structure of the genreis highly stylized and mutually agreed upon within the discourse community.Because the speaker or writer's general goal is clear when producing a genre,the problem of inability to identify speaker intention that has bedeviled ILPresearchers focusing only on free conversation can be greatly ameliorated.

A second advantage conveyed by use of the construct of discoursecommunity is the notion that such communities contain expert members:members who know the content of the field and the conventions of the discoursebetter than others. Such experts are usually easily identified by other membersof the community. The primary focus of genre analysis researchers in ESP hasbeen upon description of genres as used by expert members of the discoursecommunity. As outsiders to the discourse community, researchers themselves donot have the professional competence to understand the discourse; they need toconsult with expert members, just as any anthropologist or ethnographer mustfind a way to consult with, and describe the perspective of, members of thecommunity being studied. Selinker (1979) laid out the reasons why ESPresearchers need the input of such "subject specialist informants," and set outguidelines for working with them as informants. Researchers, he suggested,need to have informants because they themselves, as outsiders to the discoursecommunity, do not have good intuitions about the genre. Researchers shouldseek informants who are known by other discourse community members to begood at producing the genre and who in addition have the ability to reflectconsciously and analytically about the language they use in producing the genre.Recall that there are no native speakers of ESP genres: Because expert specialist

162 Tarone

informants have themselves learned the rules of their genres relatively late inlife, their knowledge of these rules seems to be more accessible to analysis andreflection than is the case with earlier-acquired native languages (see Preston,2000, for a discussion of the relative automaticity of early and late acquiredlanguage varieties). Even so, some informants are better able than others to talkwith the researcher about exemplars of the genre, providing reasons why thelanguage in the genre is organized and structured the way it is.

Genre analyses such as those described in Gibbs (2002; chapter 7, thisvolume) and Kuehn and Tarone (2000) would have been impossible without theinput of expert members of the discourse community. Basically, what expertmembers of the discourse community can do is to lay out the prototypicalstructure of given genres that are used by the discourse community. They can actas informants, showing the researcher prototypical choices made in exemplars ofthe genre. For example, when we were trying to describe the grammatical/rhetorical structure of journal articles in astrophysics, a subject specialistinformant told us that research paper writers in this field choose the passiverather than the active voice in a systematic and principled way, using the passivefor verbs referencing standard procedures and the active for verbs referencingunique choices made by the researcher (cf. Tarone, Gillette, Dwyer, & Icke,1998). Our informant glossed every verb phrase in two astrophysics journalarticles, showing us how every choice of active or passive followed this patternin the information structure of those articles.

But expert members of the discourse community can do more than justindicate what is prototypical about a genre; they can also tell us when individualspeakers or writers make idiosyncratic choices, deliberately violating theprototypical structure of a genre for some individual purpose such as languageplay (cf. Myers, 1989). An expert member can also indicate when the norms ofthe genre themselves appear to be changing. For example, Fernando's (2001)informant in plant genetics could point to cases where a younger colleaguewriting a journal article chose to use language that was more directly critical ofothers' work than the expert member himself would have chosen. Hecommented that this was an example of a growing trend among some membersof his discourse community, not one he himself participated in, but one which herecognized as part of a change in the genre.

The construct of discourse community in research on interlanguagepragmatics can also be helpful in that it naturally shifts the focus of analysisaway from the "idealized native speaker" of an idealized target language, andallows us to analyze instead the actual performance and interpretation of expertmembers of real discourse communities. Such experts may as easily benormative speakers of English as native speakers. The focus is on expertisewithin the discourse community and not on native speaker background. This isimportant because in today's world, given the increase in World Englishes, andthe use of English among normative speakers of English worldwide, it seemsless and less relevant to focus solely on the native speaker variety as the goal oflearning.

ESP and ILP 163

A final factor favoring a focus on language use inside discoursecommunities has to do with the goal of English language learners worldwide.More and more, learners are not motivated so much to learn English to becomemembers of U.S. or British society (a traditional view of integrative motivation)as to join professional discourse communities in which English is the languageof choice. If integrative motivation exists, it is increasingly focused on the desireto join discourse communities, rather than the desire to change citizenship;English language learners want to be accepted into the professional discoursecommunity of doctors, or lawyers or engineers. If this is the case, then theirinteractions within these discourse communities are the logical locus of researchon interlanguage pragmatics, because these are the settings in which they willnaturally be most focused on using and learning their new second language.

Let us now turn to some examples of studies in English for SpecificPurposes that use the constructs of discourse community and genre to analyzeinterlanguage pragmatics in natural settings.

EXAMPLES OF STUDIES IN INTERLANGUAGE PRAGMATICSUSING ESP CONSTRUCTS

Social Services Intake Interviews

The social services oral intake interview was identified by Kuehn and Tarone(2000) as a genre that is used by social service workers to screen applicants forsocial services. In this setting, the discourse community consists primarily ofsocial services financial workers, who have the common goal of identifyingindividuals entitled to social services benefits, and distributing those benefits toentitled applicants. The community may also include the applicants themselves,who may begin as novices but become expert over time, but who have a veryconstrained role in the community. The purpose of the oral intake interview as agenre is to enable a financial worker (who is the interviewer) to review informa-tion given by the applicant on a written application form in order to determinewhether and how much financial assistance should be given to the applicant.The script for this genre is agreed upon by the community as the most effectivecommunicative means to that goal, and involves the performance, by theinterviewer and the applicant, of a script: a predictable sequence of speech acts.The language used by the participants—the lexis and the structure of the scriptitself—is not easily understood by outsiders or novices, yet within the discoursecommunity these are assumed to be common knowledge.

In Kuehn and Tarone (2000), the first author had been a social worker in theoffice studied, and provided her own script (expected sequences of speech actsand typical problems) for the intake interview. In addition, two native speakerintake interviewers were asked to provide their scripts for what typicallyhappens during the intake interview. The scripts provided by the twointerviewers and the first author were the same, as one would expect, given thetheoretical framework of genre analysis which posits a discourse community

164 Tarone

with substantial agreement among its expert members on the purpose andstructure of its genres. The information structure of the written application formwas analyzed. Then, three oral interviews were audio taped, two of them withnative speaker applicants, and one with a Hispanic applicant and her daughter,who acted as an interpreter. The interviews were transcribed and analyzed.Tapes of the three interviews were played to the two intake interviewers forcomment.

The paper draws several conclusions. First, the second language learners'script includes a section that is missing in the interviewers' scripts, namely,giving reasons why she was applying for assistance. The financial aid workerssaid in the later interviews that this was a common occurrence with noviceapplicants, whether native speakers or normative speakers of English. Novicescommonly wanted to provide this information, but the financial workers did notwant or need it. Second, a common cycle in the script involves the financialworkers' paraphrase of information on the application form, with the expectationthat the applicants will explicitly confirm their paraphrase. The normativespeaker of English does not explicitly confirm this to the extent the nativespeaker applicants did. Third, there is evidence, in the conversation in Spanishbetween the applicant and her daughter, that the normative applicant oftenmisunderstands the financial worker's requests for confirmation. Possibly theydo not understand the paraphrases provided by the financial workers. Theapplicant and her daughter clearly also misunderstand some directives; they donot respond to them, discuss them in Spanish, or obey them later. The form ofthe directive "I need you to. . . ." seems particularly hard for the NNSs tounderstand. Finally, the financial workers are observed to use a large number oftechnical words and phrases (e.g., "proofs," "go retroactive," "retromedicalcoverage," "SSI," "RSDI") without explaining them and also to use slang("hunky-dory") which normative speakers of English (or even young nativespeakers) are unlikely to understand.

The authors point out the parallels between the interactional context of thewelfare office and that of the university office advising session (Bardovi-Harlig& Hartford, 1996). Both the office advising session and the oral intake intervieware genres performed in private within the discourse community, and can bethought of as private encounters of individuals of unequal status in institutionalsettings (cf. Bardovi-Harlig & Hartford, 1996). It is virtually impossible undersuch circumstances for novices to access good input ahead of time as to theappropriate L2 performance in such settings. Thus, they must learn the genre asthey take part in it, responding in interaction with a more powerful, high-statusinterlocutor. A similar context may be the doctor's office.

Doctor-Patient Interviews

Ranney (1992) examines the cross-cultural pragmatics of the doctor-patientoffice interview using the construct of "script" to analyze the genre. In thisstudy, the genre is the doctor-patient interview, an oral interaction whose

ESP and ILP 165

purpose is to enable the doctor to obtain enough information from the patient toassist him or her to reach a diagnosis of the problem which brought the patientto the office, and possibly to prescribe treatment or medication. The discoursecommunity has members that include doctors, nurse practitioners, nurses andother medical workers who assist the doctor in orienting to the problem thatbrings the patient to the office. Doctors are the expert members of thiscommunity. Patients are typically novice members of the discourse community,but if they participate frequently in the doctor—patient interview they mayachieve increasingly expert status, just as the welfare office applicants did.

Ranney's study uses the construct of script to collect data and analyze thepragmatic structure of the genre. Her data collection methodology involves theelicitation of learner and expert scripts. She proposes that Schank and Abelson's(1977) notion of script be extended beyond an expected sequence of actions toinclude an individual's expectation of a sequence of speech acts that typicallyoccur in a particular cultural event or speech event. So, for example, we all havea script, or set of expectations, based on our previous experience, about whattypically happens and is said when one enters a restaurant and orders a meal,when one buys stamps at the post office, or when one meets with one'sacademic advisor to seek approval for a proposed class schedule. Because thesescripts are based on previous experience, they are culture-bound. And becausethey are basically a set of expectations about what is likely to happen, and whatspeech acts are likely to occur in what order, scripts can be elicited in much thesame way speech acts can be elicited by Discourse Completion Tasks (DCTs).The advantage of script elicitation is that it enables the researcher to discover thelearners' expectations of discourse patterns: their ideas as to which speech actsare going to be appropriate for particular settings and roles and their under-standing of the order speech acts generally follow in a given speech event.

Using a script framework, Ranney compares college students' scripts for thegenre of the doctor-patient interview. Nine were native speakers (NSs) ofEnglish and nine were NSs of Hmong and learners of English L2. She found thatthe NNS scripts were less well developed than the NS scripts. In addition, therewere cultural differences between the scripts the two groups had for the medicalconsultation. For example, the American patients were more likely to expect tobe given a diagnosis at the end of the interview, while the Hmong patients didnot expect this. On the other hand, the Hmong patients were more likely toexpect to be given medication at the end of the consultation. Although theyconsidered it important to use respectful verbal and nonverbal behavior in thepresence of a doctor, they were also much more likely to use on-recordstrategies for requesting medication or refusing surgery than the Americanpatients were. It is unclear whether this Hmong preference for on-recordstrategies was because of their presupposition that medication was to beexpected, or because they lacked linguistic competence to use more indirectstrategies such as questions, modals such as would, or complex sentenceembeddings in making possibly face-threatening requests.

166 Tarone

Nursing Notes

The discourse community in this study consists of nurses working in an English-medium hospital in Montreal, Canada. The genre of interest is "nursing notes":documents for the patient's file that record essential information on the patient'scondition and actions taken by nursing staff. The francophone nursing studentsin the Parks and Maguire (1999) study were fortunate to have an ample supplyof mentors and coworkers in the discourse community who worked with thestudents collaboratively over time in a variety of ways to help them learn how toproduce nursing notes in English.

Parks and Maguire point out that genre analysis has traditionally focused onthe product, or the nature of the genre, but can also focus on the process oflearning how to use a new genre. They offer a longitudinal case study of afrancophone nursing student learning to produce nursing notes in English in aCanadian hospital. The study focuses on how this novice acquires the new genreon the job through the supportive mentoring of experts in the discoursecommunity. An interesting feature of this study is the inclusion in the publishedpaper of one francophone learner's hand-written first, second, and third drafts ofnursing notes, progressively incorporating more and more pieces of inputobtained from these mentors and coworkers, who told him what he had toinclude, in what order, and how it should be phrased.

Hotel Housekeeper Call-Ins

In chapter 7 of this volume, Gibbs (see also Gibbs, 2002) examines the difficultythat hotel workers have in making a particular kind of telephone call on the job.In her study, the discourse community consists of hotel workers in twodepartments: Housekeeping, and Room Service/Convention Services. Thisdiscourse community had expert members, who trained novice members on aregular basis. The genre under study is referred to in the discourse community asa "call-in." As Gibbs explains, a call-in is a phone call made by a housekeeper toa worker in Room Service/Convention Services to perform a request for removalof refrigerators, trays, roll-away beds and other items. The prototypical purposeof the call-in is thus to perform a Request speech act.

Gibbs taped call-ins made by a novice normative speaker of English and byan expert housekeeper. The fact that the call-in is a genre with a clearlyidentified goal and a prototypical structure allows us to compare the naturally-occurring pragmatic performance of the expert and novice housekeepers. In (1)below, we see an example of the expert housekeeper's call-in:

(1) "Call-in" performed by expert housekeeper:

T-O Ring.T-l R: Hello.T-2 H: Pick up in room 936, refrigerator.

ESP and ILP 167

T-3 R:936.T-4 H: Yep. (Hangs up)

In (2) below we see an example of a call-in made by the non-native speak-ing novice housekeeper, a call-in that results in pragmatic failure:

(2) Failed "call-in" performed by novice NNS housekeeper:

T-O Ring.T-l R: Hello. Convention Services.T-2 H: This is J. I'm- (.8) [housekeeping]T-3 R: [how can]

How can I help you?T-4 H: I am housekeeping.T-5 R: To call housekeeping you need to dial 52.T-6 H: Housekeeping. Pick up.T-7 R: Please dial 52.T-8 H: Pick up. Room 1717.T-9 R: I will call housekeeping to come help you. Good bye. (hangs up)

Gibbs contacted the Convention Services worker immediately after the call-in transcribed in (2) to ask why he had treated a housekeeper so poorly. Heresponded that he hadn't had any phone calls from any housekeeper all morning.

So let us compare the structure of the two call-ins. In the successful call-inin (1), the topic is introduced without any personal identification of caller orreceiver, and there is no greeting sequence, as is normally the case in socialphone calls (cf. Schegloff, 1986). In successful call-ins, housekeepers do notidentify themselves. The request is performed bald on record with no mitigation.Gibbs shows that the novice hotel worker's failure with call-ins is a result oferroneously using the structure of the social phone call, trying to providepersonal identification in Turn 2. When he does this, he is identified in as aguest by the worker in Convention Services, who responds to him as a guest inTurn 3. In spite of numerous attempts to self-identify as a housekeeper in subse-quent turns, the novice housekeeper is unable to do so. For this reason, he isunsuccessful in achieving his work-related goal of Requesting. (It is interestingthat the learners' difficulty was compounded by a hotel training tape [describedin Gibbs, 2002] that actually trained them to make call-ins the wrong way—thatis, it modeled call-ins that included personal identifications and greetings.)Gibbs concludes by showing that when the novice nonnative speaker hotelhousekeepers were taught the correct structure of the call-in, they achievedsuccess in being correctly identified in call-ins and in achieving their work-related goals.

168 Tarone

CONCLUSION

Research in ESP can enrich work in interlanguage pragmatics in a number ofways. First, in our search for research methodologies for future studies ofinterlanguage pragmatics, ESP research can provide theoretical constructs, andmethodological ideas. For example, the theoretical framework undergirdinggenre analysis, which posits the existence of discourse communities withcommon goals, shared communicative genres, and expert members who allshare knowledge about the community's discourse conventions, is very useful.Operating within the assumptions of genre analysis, the researcher can elicit anddescribe the discourse knowledge of one or two expert members of a givencommunity with some confidence that their knowledge is truly representative ofthe knowledge base of the whole discourse community. We have seen thattechniques such as script elicitation and genre analysis can be useful researchtools in describing the pragmatic competence of such informants. Such expertmembers can tell us not only what the prototypical structure of a genre in thatcommunity is, but also tell us when individual writers or speakers depart fromthat prototypical structure, and what such departures communicate.

Because ESP as a subfield generates descriptions of the way language isactually used in specified discourse communities, it can provide researchers onIL pragmatics with research methodologies and with data on the way in whichsecond language learners actually perform pragmatic functions in the real world,using the language for real purposes. In this sense, we have seen how ESPstudies have documented the way learners acquire L2 scripts and speech acts onthe job or in institutional settings. ESP studies provide data on learnerperformance in relatively high-stakes social contexts, such as academic advisingsessions, applications to college, and applications for social services.

Another trait of ESP as a subfield is its strong emphasis on establishing thegoal of learning empirically, with baseline information on the performance andjudgment of expert users of the genre in the target discourse community. Thus,ESP as a subfield can bring to the discussion on IL pragmatics information bothon prototypical and idiosyncratic expert language use, and on novice languageuse in the same genres, either to support existing theories of interlanguagepragmatics or to challenge them.

A final advantage of cultivating a connection between ILP and ESPresearch is the fact that the ESP framework provides a view of learners asparticipants in the social/institutional event/practice rather than as merelynonnative speakers. What matters in this framework is not the native vs. non-native distinction, but rather the expert vs. novice distinction. Avoidance of anoveremphasis on the native/normative distinction is also a very powerful resultof adopting the ESP-genre analysis framework.

The studies in the Appendix are selected to provide a representative cross-sampling of genre analyses that include both oral and written data produced byboth expert and second-language learning novice members of a range ofdiscourse communities. The subfield of applied linguistics that is called English

ESP and ILP 169

for Specific Purposes is extremely varied and rich; however, other studies ofinterest to ILP researchers undoubtedly exist. I hope to have provided suchresearchers with a useful starting place in searching for studies of interest tothem. Further, I hope to have provided a good case for the argument that thetheoretical frameworks and research methodologies being used in ESP are bothuseful and interesting to ILP researchers, and can help to extend the range,usefulness, and sophistication of research on interlanguage pragmatics.

ACKNOWLEDGMENT: Paper presented at AAAL '02, Salt Lake City Colloquium on RevisioningInterlanguage Pragmatics Research (A. D. Cohen, Organizer), and at the University of Illinois,Champaign-Urbana, Nov. 14, 2002.

REFERENCES

Bardovi-Harlig, K., & Hartford, B. (1996). Input in an institutional setting. Studies in SecondLanguage Acquisition, 18, 171-188.

Beebe, L., & Takahashi, T. (1989). Sociolinguistic variation in face-threatening speech acts:Chastisement and disagreement. In M. Eisenstein (Ed.), The dynamic interlanguage: Empiricalstudies in second language variation (pp. 199-218). New York: Plenum.

Broner, M., & Tarone, E. (2001). Is it fun? Language play in a fifth grade Spanish immersionclassroom. Modern Language Journal, 85, 363-379.

Brown, P., & Levinson, S. (1978). Politeness: Some universals in language usage. London:Cambridge University Press.

Eades, D. (1994). A case of communicative clash: Aboriginal English and the legal system. In I.Gibbons (Ed.), Language and the law (pp. 234-264). London: Longman.

Flowerdew, J., & Miller, L. (1996). Lectures in a second language: Notes towards a culturalgrammar. English for Specific Purposes, 15, 121-140.

Fernando, M. M. (2001). Use of passive, stative and active verb forms in two plant genetics researchpublications. Unpublished masters qualifying paper, University of Minnesota, Minneapolis.

Gibbs, T. (2002). Misidentification of limited proficiency English speaking employees in hotel call-ins. Unpublished masters qualifying Paper, University of Minnesota.

Gimenez, J. C. (2001). Ethnographic observations in cross-cultural business negotiations betweennon-native speakers of English: An exploratory study. English for Specific Purposes, 20, 169-193.

Hymes, D. (1972). On communicative competence. In J. Pride & J. Holmes (Eds.), Sociolinguistics.Harmondsworth, England: Penguin.

Jacobson, Wayne. (1986). An assessment of the communication needs of non-native speakers ofEnglish in an undergraduate physics lab. English for Specific Purposes, 5,173-188.

Kuehn, K., & Tarone, E. (2000). Negotiating the social services oral intake interview, TESOLQuarterly, 34, 99-126.

Maier, P. (1992). Politeness strategies in business letters by native and non-native English speakers.English for Specific Purposes, 11, 189-206.

Maley, Y. (1994). The language of the law. In J. Gibbons (Ed.), Language and the law (pp. 11-50).London: Longman.

Myers, G. (1989) The pragmatics of politeness in scientific articles. Applied Linguistics, 10, 1-35.Nelms, J. (2002, April). The role of sarcasm in NS-NNS (mis)-communication. Paper presented at

the annual convention of the American Association for Applied Linguistics, Salt Lake City,UT.

Nishiyama, T. (1990). A study of class discussion and participation skills. Unpublished masters PlanB Paper, University of Minnesota.

Parks, S., & Maguire, M. (1999). Coping with on-the-job writing in ESL: A constructivist-semioticperspective. Language Learning, 49, 143-175.

170 Tarone

Preston, D. (1989). Sociolinguistics and second language acquisition. Oxford, England: BasilBlackwell.

Preston, D. (2000). Three kinds of sociolinguistics and SLA: A psycholinguistic perspective. In B.Swierzbin, M. Anderson, C. Klee, & E. Tarone (Eds.), Social and cognitive factors in secondlanguage acquisition: Selected proceedings of the 1999 Second Language Research Forum(pp. 3-30). Somerville, MA: Cascadilla Press.

Ranney, S. (1992). Learning a new script: An exploration of sociolinguistic competence. AppliedLinguistics, 13, 25-50.

Rundquist, S. (1990). Flouting Grice's maxims. Unpublished doctoral dissertation, University ofMinnesota.

Schank, R., & Abelson, R. (1977). Scripts, plans, goals, and understanding: An inquiry into humanknowledge structures. Hillsdale, NJ: Lawrence Erlbaum Associates.

Schmidt, M. (1981) Needs assessment in English for specific purposes: The case study. In L.Selinker, E. Tarone, & V. Hanzeli (Eds.), English for academic and technical purposes:Studies in honor of Louis Trimble (pp. 199-210). Rowley, MA: Newbury House Publishers.

Selinker, L. (1979). On the use of informants in discourse analysis and "language for specificpurposes". International Review of Applied Linguistics, 17, 189-215.

Selinker, L., & Douglas, D. (1985). Wrestling with 'context' in interlanguage theory. AppliedLinguistics, 6, 190-204.

Swales, J. M. (1990). Genre analysis: English in academic and research settings. Cambridge,England: Cambridge University Press.

Tarone, E. (2000). Getting serious about language play: Language play, interlanguage variation andsecond language acquisition. In B. Swierzbin, F. Morris, M. Anderson, C. Klee, & E. Tarone(Eds.), Social and cognitive factors in SLA: Proceedings of the 1999 Second LanguageResearch Forum (pp. 31-54). Somerville, MA: Cascadilla Press.

Tarone, E., Gillette, S., Dwyer, S., & Icke, V. (1998). On the use of the passive and active voice inastrophysics journal papers: With extensions to other languages and other fields. English forSpecific Purposes, 17, 113-132.

Widdowson, H. G. (1998). Communication and community: The pragmatics of ESP. English forSpecific Purposes, 17, 3-14.

Willing, K. (1992). Talking it through: Clarification and problem-solving in professional work.Sydney, Australia: National Centre for English Language Teaching and Research, MacquarieUniversity.

Yule, G. (1996). Pragmatics. Oxford, England: Oxford University Press.

ESP and ILP 171

APPENDIX

Selected Annotated Bibliography of ESP Studies in Interlanguage Pragmatics

Eades, D. (1994). A case of communicative clash: Aboriginal English and the legal system. In J,Gibbons (Ed.), Language and the law (pp. 234-264). London: Longman.

Analyzes Australian Aborigines' use of a range of speech acts in English, and the ways inwhich their pragmatic practices disadvantage them in Australian courtrooms. In Aboriginalconversations, for example, rather than ask directly for information, the questioner presents aproposition for confirmation or correction. When Aboriginals seek substantial information, such asimportant personal details or reasons, they do not use questions; rather, the information-seekingprocess must be indirect and requires that the information-seeker contribute some of their ownknowledge on the topic, followed by silence. Other characteristics of Aboriginal pragmatics thataffect interaction in the courtroom include use of silence in conversation, preferred ways ofresponding to either-or questions, and difference in eye-contact patterning.

Flowerdew, J., & Miller, L. (1996). Lectures in a second language: Notes towards a culturalgrammar. English for Specific Purposes, 15, 121-140.

An ethnographic research study of lectures produced by ten native speakers of English at auniversity in Hong Kong, and the way in which L1 Cantonese speakers understood those lectures.Data were collected over a period of 3 years, via questionnaires, in-depth interviews with thelecturers and students, participant observation of 11 lectures, reflective diaries kept by lecturers andstudents, field notes, recordings and transcriptions of three lectures. The study showed that lecturersand students had very different assumptions as to the purpose of lectures: to deliver facts vs. todevelop students' judgment and thinking skills. Their views also differed as to the role of lecturers,acceptable lecture style, permissible listener behavior during lectures, and the role of humor in thelecture. Most of the lecturers' attempts to simplify their language were generally viewed by theirstudents as ineffective.

Gibbs, T. (2002). Misidentification of limited proficiency English speaking employees in hotel call-ins. Unpublished masters qualifying paper in Linguistics, University of Minnesota.

The researcher describes the conversational structure of the "call-in", a type of phone call madeby hotel employees to other hotel staff. The function of the call-in is to perform a Request speechact. Call-ins made by native speaker and non-native speaker hotel staff are taped and described.Native speaker call-ins are shown to have a quite different structure from that of a social phone call:call-ins lack a greeting sequence and a personal identification sequence. This difference isproblematic for nonnative speakers, whose training materials compound the problem.Communication breakdown occurs when NNS hotel staff try to make call-ins using theconversational structure of social phone calls and are misidentified as hotel guests. Training in theactual structure of the call-in results in successful performance of the Request speech act by LEPhotel employees.

Gimenez, J. C. (2001). Ethnographic observations in cross-cultural business negotiations betweennon-native speakers of English: An exploratory study. English for Specific Purposes, 20, 169-193.

The researcher analyzes NNS-NNS negotiations in business interactions and finds that sellersmake more Commitments and Promises, using high self-disclosure, while Buyers perform moreWarning speech acts (e.g., warning that they may go to competitors). Bargaining style can be eithermonochrome (a linear temporal orientation, with a rational decision criterion) or polychrome (anindirect, circular approach, with an intuitive decision criterion). The researcher analyzes theperformance of "Reject" and "Suggest" speech acts; other important moves are "Establishingcredentials", "Negotiating prices". Cross-cultural differences in strategy use are analyzed.

172 Tarone

Jacobson, W. (1986) An assessment of the communication needs of non-native speakers of English inan undergraduate physics lab. English for Specific Purposes, 5, 173-188.

The researcher audiotaped students engaged in pair work in an undergraduate physics lab at theUniversity of Minnesota. The oral interactions of native-speaker pairs are compared to theinteractions of normative speaker pairs. Much more information was exchanged orally when thepartners were both NSs than when one or both were NNSs. The communicative failure of a NNSstudent attempting to perform a Request for assistance from a TA is analyzed. In comparison to NSRequests, this student's Request was not sufficiently specific. It did not include enough backgroundinformation about attempts to solve the problem that had already been tried. The NNS responded toTA inquiries with only affirmative or negative signals. As a result, the TA was unable to assist himin a timely fashion, spending more than 20 minutes with him during the lab period.

Kuehn, K. and Tarone, E. (2000). Negotiating the social services oral intake interview, TESOLQuarterly, 34, 99-126.

The study describes the script followed in the oral intake interview of a social services office,showing how it is related to the printed application form the applicant had filled out. A NNSapplicant is shown to have a different script from the interviewers, to provide minimal back channelcues, and to misunderstand such speech acts as Confirmation Requests and Directives made by theinterviewers.

Maier, P. (1992). Politeness strategies in business letters by native and non-native English speakers.English for Specific Purposes, 11, 189-206.

The researcher asked native and nonnative English speakers to write business letters inresponse to a role-play situation. In this situation, the writers have missed a job interview due totransportation problems, and have to write a letter requesting a new appointment for an interview.Native and non-native writers are shown to use quite different politeness strategies: the NNSs tendedto use more positive politeness strategies, stressing their strong desire to secure the position and theiroptimism that they would be able to obtain a second interview, while the NSs tended to use negativepoliteness strategies, apologizing profusely for the imposition of having missed the firstappointment, and stressing the interviewer's freedom to deny them a second interview.

Myers, G. (1989). The pragmatics of politeness in scientific articles. Applied Linguistics, 10, 1-35.

Myers shows how the Brown and Levinson (1.978) model of politeness (previously appliedonly to oral discourse) can elegantly account for a wide range of seemingly unrelated phenomena inthe written genre of research articles—phenomena such as hedging in making claims, affectation ofmodesty in referring to one's contributions, patterns of citation (or not) of others' work, and the useof humor in naming phenomena. Myers suggests that we view the research article as a delicatebalancing act between the opposing pragmatic forces of performing the face-threatening act ofmaking a claim, while mitigating that act. To mitigate, the writer uses positive and negativepoliteness strategies to preserve the faces of colleagues in the discourse community who may bethreatened by that claim. Negative strategies include use of hedging devices such as modals inmaking claims. Positive strategies include expressions of dismay at the limitations of positionsthreatened by the claim, and the use of language play in the form of humor and puns.

Nishiyama, T. (1990). A Study of class discussion and participation skills. Unpublished masters,plan B paper, University of Minnesota.

The researcher audiotaped one hour of a class session of a graduate course, MBA 8045(Marketing Management) that used the case study approach. This technique is based on classdiscussion and group decision-making. The researcher also analyzed questionnaires returned by 21of the 42 students in the class, of whom 5 were normative speakers of English. Unlike the NSs, theNNSs indicated class discussions were the most problematic aspect of the course. The researcheridentifies 3 types of interaction in the class, and analyzes the differing roles of the teacher in those

ESP and ILP 173

interactions. She concludes that NNSs need to understand the way the teacher uses a range ofconversational moves to signal to the students what is and is not relevant to the discussion.

Parks, S., & Maguire, M. (1999). Coping with on-the-job writing in ESL: A constructivist-semioticperspective. Language Learning, 49, 143-175.

This longitudinal case study of a francophone nursing student learning to produce nursingnotes in English focuses on how a novice acquires a new genre on the job through supportivementoring of experts in the discourse community. Three consecutive drafts of the student's notes ona case are shown, together with the revisions that were suggested at each stage by more experiencednurses. The study shows how experts' suggestions are incorporated in each subsequent learner draft.

Ranney, S. (1992). Learning a new script: An exploration of sociolinguistic competence. AppliedLinguistics, 13, 25-50.

This study compares scripts for the medical interview of 9 NSs of English and 9 NSs ofHmong and learners of English L2 (all college level). Nonnative speakers had systematicallydifferent scripts for what they expected to take place and what they expected to be said in a doctor-patient office interview. For example, Americans expected to be given a diagnosis at the end of theinterview, while the Hmong expected to be given medication at the conclusion. NSs were moreindirect than the NNSs in making requests.

Schmidt, M. (1981) Needs assessment in English for specific purposes: The case study. In L.Selinker, E. Tarone, & V. Hanzeli (Eds.), English for academic and technical purposes:Studies in honor of Louis Trimble (pp. 199-210). Rowley, MA: Newbury House Publishers.

The lecture comprehension and note-taking strategies of an Asian business student areanalyzed in a college-level business class. The student's strategy of taking notes in Chinese is shownto backfire when she has to take an essay exam in English, because she doesn't have a record of theEnglish language discourse needed to display what she knows on the exam in English. Though sheaccurately copies down a table written on the board, she is unable to explain the relationships itrepresents in English (a problem of information transfer, translating information from a visual formto words, or vice versa).

Willing, K. (1992). Talking it through: Clarification and problem-solving in professional work.Sydney, Australia: National Centre for English Language Teaching and Research, MacquarieUniversity.

The researcher integrates frameworks of speech act analysis, conversation analysis, scriptanalysis to analyze the interactive signaling used by native and nonnative speakers of English as theywork together on problems and tasks onsite in white collar professional, multicultural workplaces.Participants are equipped with pocket tape recorders, which they switch on when engaged inconversational problem solving with their coworkers in dyads and small groups. The researcherdescribes clarification strategies deployed in dealing with cross-cultural and pragmatic communi-cation difficulties.

7

Using Moves in the Opening Sequence toIdentify Callers in Institutional Settings

Tara Leigh GibbsUniversity of Minnesota

Students often come to teachers with questions about what to say in a particularsituation. It is tempting to answer these questions with a spontaneously createddialogue and some appropriate vocabulary or syntax. Rarely, however, does theanswer describe the various pragmatic moves the addressee is expecting or howthe addressee will interpret deviations in the moves. Yet research suggests thatthe greatest source of difficulty in communication may not come from syntaxand phonology, but rather from a failure to meet the pragmatic expectations ofthe discourse community.

This chapter examines the pragmatic structure of a particular type ofinstitutional phone call, known as a call-in, performed daily by staff at a largeconference hotel. After describing the setting for this research, this chapterreviews the structure of the opening of social telephone conversations identifiedby Schegloff (1986). Then, it describes differences between social andinstitutional phone conversations discovered by Hopper, Doany, Johnson, andDrummond (1990) and Wakin and Zimmerman (1999). Next, it describes themethodology employed in this study and analyzes the structure of a call-in by anexpert member of the hotel's discourse community and the structure of call-insby unsuccessful novice speakers. Finally, it compares these with the previousresearch described at the beginning of the chapter.

The Research Setting

Call-ins are phone calls in which employees, especially housekeepers, requestservices from another department, services such as the pick-up from a guestroom of a tray, roll-away bed, or refrigerator, or the repair of something bymaintenance. At least three largely unsuccessful attempts were made by thehotel to train its Limited Proficiency English Speaking (LPES) employees toperform call-ins prior to the ESL course from which this research materialderives. The three attempts included two ESL programs prior to 1999 and avideo created by the hotel and shown during new employee orientations.

In 1998, an ESL program which included a job shadowing component wasinitiated. In job shadowing the ESL teacher follows workplace ESL studentsone-on-one while they are working in order to encourage the students to use theEnglish they are practicing in the classroom and to learn what types of language

175

176 Gibbs

they are struggling with on the job. Job shadowing often included videotaping,audiotaping, or photographing in order to produce authentic material for thestudents to practice with in the classroom. The telephone conversations analyzedin this chapter were collected by me, the ESL teacher, during job shadowing andwere used by the students in their class in order to become more aware of errorswhich led to unsuccessful call-ins and the nuances of the task they were trying toperform.

The results, presented below, were bewildering at first, especially whenviewed in light of the familiar social phone call, and suggested a great deal ofinsensitivity on the part of the Convention Services employees with whom theyinteracted; however, this did not jive with what I knew of the ConventionServices employees. The root of the problem became more apparent when aConvention Services employee was interviewed immediately after anunsuccessful call-in was recorded and said, "But, no one from Housekeeping hascalled all morning." He had misidentified the caller as a guest rather than as anemployee.

Various learning challenges, such as, "Why aren't LPES employeessuccessful when making call-ins?" and "How can we help the LPES employeesto make call-ins?" did not seem answerable simply by watching what theworkplace ESL students did while working. As a result, the concept of jobshadowing came to be expanded in this program to also include shadowingexpert speakers (native or nonnative speaking supervisors and coworkers whowere successful in their communicative tasks) to find out how they performedtheir jobs.

Shadowing expert speakers at their jobs, in addition to shadowing thenovice speakers, allowed differences in the pragmatic expectations of the expertspeakers and the novice speakers to become apparent. This in turn allowed the1998 ESL program to educate workplace students on the correct call-instructure, rather than on creative, but incorrect, structures which resulted inunsuccessful call-ins.

BACKGROUND AND LITERATURE REVIEW

Social Telephone Calls

Schegloff (1968, 1979) studied 450 telephone calls in a variety of settings,including personal phone calls and phone calls to businesses. Excluding allphone calls that did not begin with a "hello-hello" sequence (e.g., "AmericanAirlines, how can I help you?"), Schegloff (1986) identified a series ofadjacency pairs (p. 117) that almost invariably occur at the beginning of socialtelephone calls and that always impute a special meaning, for example,irritation, when they do not occur.1 This sequence, known as the opening

1 Adjacency pairs are sequences in which the first part obliges the second part, for example, aquestion and an answer, a greeting and a response, a display for identification and a display that

Using Moves in the Opening Sequence to Identify Callers 177

sequence, is composed of a summons-answer sequence followed by anidentification sequence followed by a greeting-return greeting sequence,followed by a how-are-you sequence. Only after this opening sequence does thecaller typically introduce the topic. Example (1) illustrates a minimal openingwithout a how-are-you sequence (Levinson, 1983, p. 312).2

(1) Canonical opening sequence

C: ((rings)) ((SUMMONS))Tl R: Hello. ((ANSWER))

((DISPLAY FOR RECOGNITION))T2 C: Hi. ((GREETINGS 1ST PART))

((CLAIM THAT C HAS RECOGNIZED R))((CLAIM THAT R CAN RECOGNIZE C))

T3 R: Oh, hi:: ((GREETINGS 2ND PART))((CLAIM THAT R HAS RECOGNIZED C))

Schegloff (1968), Sacks, Schegloff, and Jefferson (1974), Levinson (1983),Psathas (1995), Markee (2000) and others state that all conversation is governedlocally via two-part action sequences called adjacency pairs. The first utterancein the pair is called a first, and the second utterance in the pair is called a second.In the foregoing example, turn zero, the summons, is considered a first. Theresponsive action, the answer, in Tl, is considered the second.3 Most actions,such as the introduction of topic, have preferred seconds (acceptance of thetopic) and dispreferred seconds (rejection of the topic). A preferred second isthe most common or expected utterance in the given situation. A dispreferredsecond is an uncommon or undesired utterance in the given situation. Accordingto Sacks et al. (1974), a first always demands a second. Failure to provide onecarries special meaning and implications, while providing a dispreferred secondmay require additional rhetorical support.

Seconds often double as firsts for the next adjacency pair. In Example (2) Iindicate the adjacency pairs for Tl of Example (1). Tl, in bold, shows thisdoubling of duties, known as interlocking organization (Schegloff, 1986).

(2) Interlocking organizationC: ((rings)) 1st ((SUMMONS))

Tl R: Hello. 2nd ((ANSWER))

identification has or has not occurred, an introduction of topic and an uptake/acceptance/rejection ofthat topic, or a move to close a conversation and an agreement/rejection of that move. They arecanonically adjacent, hence the name, adjacency pair.2 Later descriptions of the Opening Sequence tend to omit the "How are you" sequence. It may bethat the "How are you" sequence is actually a presequence opening up the topic. As a result, thisportion of the canonical sequence is abridged in its description here, inasmuch as the relevant pointof the analysis is that the topic opens AFTER the identification and greeting sequences arecompleted and that this is true regardless of the interpretation of the "How are you" sequence.3 See Schegloff (1986) for an explanation of why this is an answer, not a greeting.

178 Gibbs

1st ((DISPLAY FOR RECOGNITION))

T2 C: Hi. 2nd ((CLAIM THAT C HAS RECOGNIZED R))

Sometimes, a single utterance can initiate two firsts. In Example (3) I againindicate the adjacency pairs found in Example (1) and highlight T2 which, inaddition to illustrating interlocking organization, also initiates two firsts. Inresponse, T3 replies with two seconds.

(3) Packing two firsts into a single utteranceC: ((rings)) 1st ((SUMMONS))

Tl R: Hello. 2nd ((ANSWER))

1st ((DISPLAY FOR RECOGNITION))

T2 C: Hi. 2nd ((CLAIM THAT C HAS RECOGNIZED R))

1st (a) ((CLAIM THAT R CAN RECOGNIZE C))1 lst (b) ((GREETINGS 1ST PART))

T3 R: Oh, hi:: 2nd (b) ((GREETINGS 2ND PART))2nd (a) ((CLAIM THAT R HAS RECOGNIZED C))

The compression of two firsts into a single utterance or two seconds into asingle utterance, rather than taking two turns or using two sentences to get theideas across, will be referred to as packing since I am unaware of a commonterm to describe this phenomena. A more precise definition of packing wouldbe: the occurrence of two firsts or of two seconds in one turn constructional unit.According to Levinson (1983), turn constructional units are blocks of speechthat are used to build turns in speech. The end of a turn constructional unitconstitutes a transition relevance place because the next turn can transition toanother speaker just at this point in time. A turn constructional unit may becomposed of sentences, clauses, noun phrases, or even single words. A turn mayactually be composed of several turn constructional units if the speaker refusesto yield the floor at a transition relevance place, or if the current speaker is alsoallocated the next turn constructional unit.

Packing the greeting and claim for recognition together seems fairlycommon in the telephone calls Schegloff studied. Schegloff (1968) explainedthis by noting that callers know whom they hope to reach, though notnecessarily who is answering the phone, and the receiver does not haveknowledge, initially, as to who the caller is (remembering that Schegloff s workwas done prior to caller I.D., of course). Thus, there is and must always be anidentification sequence at the beginning of a phone call, even if it is not overtlyrealized. An example of an overt realization would be someone saying, "Hi, thisis Mary. Is this Bert?" whereby the turn has three turn constructional units andthree actions—greeting, claim the caller can be recognized, and claim the caller


has recognized the receiver—as compared to T2 of Example (3) which has oneturn constructional unit and all three of these same actions.

Institutional Telephone Calls

There is evidence that the canonical opening sequence identified by Schegloffmay be altered in institutional settings. This section describes research byHopper et al. (1990) which examines a phone call to a doctor's office by apatient, and research done by Wakin and Zimmerman (1999), Whalen andZimmerman (1987, 1990), and Zimmerman (1992), which investigates phonecalls from citizens to emergency dispatchers and to telephone directoryinformation operators (henceforth referred to by their respective 3-digit phonenumbers in the United States—911 calls and 411 calls). In each of these phonecalls the interactions are between outsiders to a particular discourse community(patient or citizen) and members of the particular discourse community(receptionist or dispatcher/operator).

Hopper et al. (1990) observed that data on telephone openings in doctors'offices include more examples that deviate from the canonical opening sequencethan examples that conform to it. In light of openings such as that in Example(4), Hopper et al. (1990, p. 372) suggest that "strangers or previouslyunacquainted parties often display reduced formats compared to openingsbetween acquaintances."4 In Example (4), I have added the turn information andaction sequences and changed the names to simply C for caller and R forreceiver. The opening sequence in Example (4) displays a summons-answersequence, part of an identification sequence, and an introduction of topic. It doesnot display a greeting sequence, nor does it display a how-are-you sequence.However, T2 does show an overt answer ("yes") to the move in Tl, which Iinterpret as a claim that the caller has identified the receiver, or at least the factthat she is a receptionist at the clinic he is trying to reach. This is followed by anovert identification of self by the caller, and, after a pause, an introduction oftopic. The sequence includes a number of pauses and disfluencies, which mightbe indications of attempts to allocate turns or to give dispreferred seconds, suchas "I don't recognize you." Hopper and colleagues do not analyze this in detail,and so my suggestions regarding these disfluencies are followed by questionmarks.

4 Hopper et al. state that in conversations between strangers there are fewer presequences, se-quences that set up a move such as the introduction of topic or the closing of a conversation. (Formore information on closing sequences see Schegloff et al. (1973). However, in Example 4 itappears to me that the "U::m" in T2 might be interpreted as a presequence setting up the introductionof topic, albeit with substantially reduced overt content. Likewise, the pause of four-tenths of asecond could be a failure by the receiver to reject the move to introduce a topic and thus allocationof the next turn constructional unit to the caller, who then proceeds to introduce the topic.

180 Gibbs

(4) Opening sequence of a call to a doctor's office

TO: C:Tl: R:

T2: C:

((Ringing))Central Allergy Associatesthis is Bonnie?(0.9)

Ye:sThis is Rick Harrell(0.4)

U::m(0.4)couple a days ago I sawDoctor: uh (0.4) Hart on auh (1 .0) I was having coughing

((SUMMONS))((ANSWER))((DISPLAY FOR RECOGNITION))((Turn allocation? Claim by C not torecognize the name Bonnie?))((CLAIM OF RECOGNITION OF R))((DISPLAY FOR RECOGNITION))((Turn allocation? Claim by R not trecognize C?))((Request to initiate topic?))((Permission to initiate topic?))((INTRODUCTION OF TOPIC))

Similar findings for institutional phone calls to 411 (directory assistance)and 911 (emergency dispatchers) are reported by Whalen and Zimmerman(1987). Whalen and Zimmerman attribute such altered opening sequences to theidea that relational constraints are often not separable from the communicators'goals. Wakin and Zimmerman (1999), Whalen and Zimmerman (1990), andZimmerman (1992), document monofocal phone calls to 411 and 911 in whichcitizens are calling professionals at these phone numbers. They report theomission of three sequences from the opening sequence: the identificationsequence, the greeting sequence, and the how-are-you sequence.5 They find thatthe goals of these calls override the need for these sequences, and that thepreemption of these sequences demonstrates the orientation of the parties to thecommon goal: getting a phone number and reporting an emergency.

In Example (5), Wakin and Zimmerman (1999, p. 416) have used D forDispatcher instead of R for receiver and they have numbered the dialogue byturn constructional unit rather than by turn. I have added the actions in italics forwhich they did not offer an analysis.

(5) Opening Sequence of a 911 Call

00 ((Ring))01 D: Mid-City emergency::,

02 (0.1)03 C: U::myeah.04 Somebody just vandalized my car.

((SUMMONS))((ANSWER))((IDENTIFICATION))((??))((CLAIM OF RECOGNITION?))((INITIATION OF BUSINESS))

5 In the 911 calls identification does occur as part of the subsequent talk, albeit not usually duringthe opening. Additionally, when callers do identify themselves during the opening sequence, thedispatcher doesn't seem to orient towards this at all at this point in time and later re-requestsidentification.


Zimmerman and Wakin express concern that the monofocal nature of thesephone calls may make the structures ungeneralizable to other institutional phonecalls. When comparing the phone call to the doctor's office with the phone callto 911, we can observe differences in the identification sequence and in themoves to introduce the topic of discussion. However, there are also similarities.Both phone calls omit the greeting sequence and the how-are-you sequence, andin both cases, mutual, personal identification has not occurred by the time thetopic is introduced. A fourth, potential commonality is the disfluency precedingthe introduction of topic. Thus, there may be some characteristics that can begeneralizable to other institutional phone calls.

The calls to the doctor's office, to 411, and to 911 are all calls made bypeople who are not members of the institution's discourse community to peoplewho are members of the institution's discourse community. The data in thischapter, however, are from members of an institutional discourse communitycalling other members of the same discourse community. Fairchild (1995), whostudied politeness in institutional phone calls, suggested that it may be possibleto identify callers as clients or employees even without their identifyingthemselves overtly as such. Thus, in addition to asking whether institutionalcalls are different from social calls, we should also ask if members of theinstitutional discourse communities talk with each other the same way they talkwith outsiders to the discourse community.

Genre and Discourse Communities in Institutions

Swales (1990) defined genre as a property of a discourse community. Adiscourse community is a group of individuals who share a set of commonpublic goals, use mechanisms of intercommunication to provide information,use one or more genres to pursue their goals, use some special lexis, and includemembers with content and discourse expertise (pp. 24-27). Swales offers thefollowing five criteria for a genre:

1. A genre comprises a class of communicative events used by a discoursecommunity to further its common goals.

2. The class of events share a single communicative purpose3. The events vary in their prototypicality, but their structure is highly

predictable4. The speech event occurs frequently within the discourse community.5. The speech event is clearly recognized by experts in the discourse

community.

The goal of genre analysis seems to be the explication of the forms used in aparticular situation with a particular meaning.

The call-in appears to meet Swales' (1990) criteria for a genre. Specifically,there is a discourse community with expert members who recognize the call-inas a speech event that occurs frequently and who have sought to train others toconduct this speech event. The call-in is part of a class of communicative events

182 Gibbs

with a single goal: requesting another department to do something. Finally, thecall-in varies in its prototypicality—not every call will be word-for-word thesame, and some calls will flout the rules to various ends—but its structure ishighly predictable as this chapter will show, and expert members of thediscourse community expect and utilize this structure as members of thediscourse community in order to display themselves as employees and recognizeeach other as employees.

The genre analysis framework assumes the existence of expert members ofthe discourse community who know, understand, and use the discourse rules ofthe community, and novice members of the discourse community who may notyet know, understand, or use the discourse rules of the community (Swales,1990; Tarone, chapter 6, this volume). Examples of novices are universitystudents learning how to write academic research papers (Swales, 1990), and, asthis chapter shows, housekeepers learning to make call-ins.

RESEARCH QUESTIONS

Given our goal of understanding why the novice speakers are unsuccessful inmaking call-ins, we would like to see if there are any differences between thestructures of successful and unsuccessful call-ins which might account for thisand if, provided an accurate model, novice speakers are able to perform asuccessful call-in.

METHOD

As the ESL teacher for the course which was training students to perform call-ins, I quickly became aware that the students avoided call-ins because they were"difficult," despite their success in calling each other room-to-room duringclassroom role plays of call-ins. Since one of the principles of the curriculumdevelopment was to bring as much of the job into the classroom as possible andas much of the classroom into the job as possible, permission to record thestudents making call-ins was quickly received during informal meetings with thehotel administration. Arrangements were made during a weekly chat with thehead of housekeeping and during a daily exchange of greetings with the head ofhuman resources by the photocopier. I had hoped that listening to their own callswould help novice housekeepers build confidence, identify problem areas, andbridge the gap between role play and reality. An expert speaker was alsorecorded to provide an instructional model for the students.

Data Collection

The data analyzed in this chapter were gathered as part of the needs analysis andcurriculum development materials for an ESL class. Three types of data wereutilized in this study: transcripts of call-ins, interviews with employees, and aninformal listening survey of expert English-speaking employees.


Both expert English-speaking and limited-proficiency English-speakinghousekeepers were recorded during job shadowing. Job shadowing wasscheduled after every class for approximately 1 hour. Call-ins were recordedwith a Marantz PMD222 tape recorder that was plugged into the telephone inthe rooms being cleaned or with a Radio Shack suction-cup Telephone Pick-upplugged into a Sony TCM-353V cassette-recorder. The expert English-speakingsupervisor was recorded when I followed her on her rounds. She received nospecific instructions regarding the call.

The limited-proficiency English-speaking housekeepers were recordedtwice. A year prior to the first recording, a handful of the students had been in adifferent ESL classes and had practiced making call-ins. In addition, most of thestudents had seen a video of a call-in during the new employee orientation. Inmy class, prior to the first recording, the students had role-played making a call-in with each other. The role plays were spontaneous and unscripted. At the timeof the first recording, learners had not been given an authentic call-in dialogue touse while making call-ins, and thus, they were spontaneous and unrehearsed.

Approximately 2 weeks later the novices were recorded a second time.Before the second recording, they received instruction on making a call-in. Theinstruction consisted of providing the learners with a transcript of the expertsupervisor's call-in with blanks for the room number and item to be picked up.The students practiced calling a supervisor in another guest room using this newdialogue worksheet. During job shadowing students were encouraged to attemptmaking a real call-in to Convention Services and these calls were again recordedjust as the first calls had been recorded.

On some occasions, expert English-speaking employees who were involvedin the recorded conversations were also interviewed in order to determine whomthey believed they had been talking to. In particular, this happened in personafter the primary subject's first recording, and it happened several other timesover the phone when I called Convention Services after a student had hung upand asked them if they thought they had been talking to an employee or a guest.However, after a few calls they began to realize why I was calling, which madetheir responses less spontaneous.

In the last stage of data collection, the recording of the supervisor wasplayed for 16 expert English-speaking employees. The employees wereapproached individually and in small groups during lunchtime and askedinformally to listen to a conversation on tape and to identify the participants inthe conversation as employees of the hotel or guests of the hotel. Recordings ofthe limited-proficiency English-speaking novices were not played for expertEnglish-speaking employees in order to avoid embarrassment, since theemployees may have recognized each other's voices.

Participants

The primary novice participant is a 50+ year-old male Vietnamese LimitedProficiency English Speaking (LPES) housekeeper who had worked at the hotel

184 Gibbs

for more than 1 year and who attended a mandatory worksite ESL course 1.5hours a day, 2 days a week. He also participated in job shadowing with the ESLinstructor after each class. He was chosen as the primary subject for the analysisbecause his data was the most complete data in the curriculum archive (see thefollowing section on data collection), including a worksheet with hisconversation and a worksheet recording his pause lengths for an exercise onholding the floor in a conversation. Finally, it was after taping his call-in that Italked with the employee in Convention Services who claimed not to havespoken with any housekeepers "all morning." Thus, there was additionaltriangulation of the data regarding this conversation.

As the course progressed, I began following one of the supervisors aroundduring job shadowing in order to observe her interactions with the students. Hercalls were extremely striking in their speed, abruptness, and consistency, so Iasked her if I could record one of her calls to use as a dialogue model. Theprimary expert subject is a 50+ year-old female expert English-speakinghousekeeping supervisor who had worked at the hotel for over 20 years and whowas training a new native English-speaking housekeeper at the time of therecording. Each time she came to a Limited English Proficiency Employee andthey reported call-ins which needed to be made, she performed a call-in.

The secondary subjects were 16 expert English-speaking supervisors andexpert English-speaking employees in Housekeeping, Human Resources, RoomService and Convention Services, and 18 LPES housekeepers participating inthe ESL course and job shadowing. The LPES housekeepers had variouslanguage backgrounds, including Spanish, Amharic, Oromo, Somali, andVietnamese. Their ages ranged from 19 to 62.

RESULTS AND DISCUSSION

This section presents six transcripts of call-ins to Convention Services. The firsttranscript is of a successful call-in by the expert speaker. This is followed bythree transcripts of unsuccessful call-ins by novice speakers, and two successfulcall-ins by novice speakers. The section ends with two successful call-ins by thefirst and third unsuccessful novice speakers after they have been trained on thedialogue of the expert speaker.

Expert Speaker Call-in

Example (6) shows an expert speaker's call-in.

(6) Expert Speaker: Successful call-in

TO: C:Tl: R:

T2: C:

RingsHello,

Pick up in room936, refrigerator.

((Summons))((Answer))((Display for recognition))((Claim C has recognized R))((Claim R can recognize C))((Intro, of topic — Request))


T3: R: 936. ((Granting of Request—Uptake of Topic))((Claim R has recognized C))((Question))

T4: C: Yep. (Hangs up) ((Answer))((Closure))

T5: ((presumably R hangs ((Closure))up, too))

(Transcript of Z., recorded in May, 1998)

In TO and Tl (Turn 1) of the expert speaker's transcript, we see a first, thesummons in TO, with an answer as its second in Tl. Tl also serves as a first, adisplay for recognition. T2 provides a second, a claim the caller has recognizedthe receiver, and it provides a first, an introduction of topic as a request. Basedon Schegloff s argument that there must be identification occurring for both thereceiver and the caller, and his action-sequencing for telephone calls, anidentification sequence must have occurred prior to the introduction of topic.Because T2 is the first time that the caller speaks, the claim the receiver canrecognize the caller cannot occur earlier than T2. As the topic is clearlyintroduced in T2, this leaves only one conclusion—that T2 provides two firsts.Thus, in addition to introducing the topic—a request—T2 also makes a claimthat the receiver can identify the caller.

T3 offers a second to both of these firsts. By confirming the room number,the receiver is taking up the topic and granting the request. Additionally, basedon our acceptance of the theory that a first requires a second, we must assumethat the receiver is implicitly confirming his identification of the caller as ahousekeeper. Notice that this analysis has the introduction of topic occurringbefore the identification sequence is completed. In a subsequent interview, theman who answered the phone in Convention Services after the phone callindicated that "a housekeeper had called." An interview with the housekeeperafter the conversation indicated that one of the "young guys" had answered thephone in Convention Services. Thus, the call had successfully enabled mutualidentification, in terms of the institutional role of the caller and receiver, to takeplace.

The identification process was also verified when this phone call and threeother models were played for 16 native English speaking employees in the lunchroom from various departments, including Convention Services, Housekeeping,Maintenance, and Security. They were asked to determine whether the callerwas a guest or an employee. All identified the caller as a hotel employee.Several added that the caller was probably a housekeeper. Note that identifica-tion occurs, not in terms of individual identity, but in terms of the role of thecaller and receiver.

186 Gibbs

Structure of a Call-in by a Novice Speaker

Analyzing the novice speaker's transcript in terms of Schegloff's categoriesyields the sequences seen in Example (7). Note that in the sections in whichrepairs occur, the caller's interpretations of the actions are indicated in bold,while the receiver's interpretations of the actions are indicated in italics.

(7) First novice speaker: Unsuccessful call-in

TO:Tl:

T2:

T3:

T4:

T5:

T6:

T7:

T8:

T9:

C:R:

C:

R:

C:

R:

C:

R:

C:

R:

RingsHello.Convention Services.This is J. I'm-(.8)[housekeeping.][How can-]How can I help you?I am housekeeping.

To call housekeepingyou need to dial 52.

Housekeeping. Pick up.

Please dial 52.

Pick up. Room 1717.

I will call housekeepingto come help you.

Good-bye.

1st ((Summons))2nd ((Answer))1st ((Claim that C can recognize R))2nd ((Claim that C has recognized R))

1st ((Claim that R can recognize C))2nd ((Claim that R has recognized C))1st ((Offer/Request for intro. of topic))1st ((Repair of T3: Claim that R has

recognized C: assertion: Claim thatR can recognize C))

2nd ((Introduction of topic: Answer))1st ((Request for housekeeping))2" ((Uptake of request: Answer))1st ((Directive))2nd ((Rejection of repair: directive to

call himself))1st ((Repair of T3: Claim that R has

recognized C: assertion: Claimthat R can recognize C))

1st ((Introduction of topic: request forpick-up))

2" ((Rejection of directive to callhousekeeping))1st ((Request for housekeeping))2nd ((Uptake of request: Answer))

1st ((Directive))2nd ((Rejection of repair))2nd ((Rejection of topic))

1st ((Repair of introduction of topic:request for pick-up))

2nd ((Rejection of directive))1st ((Request for housekeeping))2nd ((Answer))1st ((Decision to circumvent pre-

closure))1st ((Closure))


T10: C: Pickup. Is- (.8) 1st ((Repair))I-(.8) ((??))There-(l sec)? ((??))

(Transcript of J. recorded in April, 1998)

In Example (7), TO is a first, a summons to the phone. Tl is a second, ananswer to the summons. It is also a first, a claim that the receiver can recognizethe caller, also called a display for recognition. T2 is a second, a claim the callerhas recognized the receiver. T2 is also a first, a claim the receiver can identifythe caller, also called a display for recognition. This display is overt. The calleralso begins a second sentence, apparently an attempt at further clarification ofhis display for recognition, however, he is unable to come up with the termhousekeeper and after a long hesitation offers the word "housekeeping" instead.

In T3 we see that since the caller has displayed himself for recognition inT2, the expectation is that the receiver will either make a claim that he hasrecognized the caller and move on to the next action sequence, or he will make aclaim that he hasn't recognized the caller in T3 and request additionalinformation. In T3, the receiver overlaps the speech of the caller and then repairsthis overlap by starting his sentence over. His claim that he has recognized thecaller is not overt, but is implied since he moves the conversation forward byrequesting the caller to introduce the topic of conversation. Please note thatregardless of the actual accuracy of the identification, T3 is still making a claimthat the receiver has recognized the caller.

In T4 the caller says, "I am housekeeping," a repair of his attempt toidentify himself as a housekeeper during his uncompleted turn in T2, which helost when he hesitated. However, based on the receiver's response in T5, thereceiver appears to interpret this as an answer to his T3 request for anintroduction of topic, in other words, the receiver sees this as a second, anintroduction of topic: an answer to his request, and also as a first, a request forhousekeeping.

In T5 the receiver, who has understood T4 as a request, offers a dispreferredsecond—an answer regarding how to reach Housekeeping, rather than a grantingof the request which would be the preferred second. From the receiver'sperspective, T5 is also a first, a directive to call housekeeping.

From the caller's perspective, a different set of actions has occurred. Thecaller views this as an answer to the repair, an answer that indicates the repairwas unsuccessful. This is evidenced by the fact that in the caller's next turn heagain attempts another repair, indicating that he realizes his first repair wasunsuccessful.

In T6 the caller again attempts a repair at identification, and then proceedsto the next action sequence—the introduction of topic, a request for a pick up.The receiver, however, again fails to interpret this as a repair of the identifica-tion sequence or as an introduction of topic: a request for a pick-up. Instead, he

188 Gibbs

seems to see it as a rejection of the dispreferred second, the directive, and as arenewed request to speak to Housekeeping.

In T7 the receiver responds to the request in T7 with a reiteration of hisanswer, a directive to dial 52. The caller realizes that the repair has again beenunsuccessful and that the receiver has not taken up the topic he has introduced.

In T8 the caller again attempts a repair, reintroducing the topic. Thereceiver, however, sees this again as a rejection of his directive and a request forhousekeeping in room 1717.

In T9 instead of offering a second, the taking up of the caller's topic whichwould have required the receiver to determine anew the caller's topic, thereceiver seems to offer a second to his own first from T8, responding to his owndirective. Another possibility is that the receiver is offering what he perceives tobe a more preferred second to the caller's apparent request for housekeeping, apromise to call housekeeping for the caller. The conversation ends abruptlywithout a preclosing sequence6 when the receiver nominates himself for asecond turn constructional unit in T9 and says "Good-bye" without consultationwith the caller.7

In T10 the caller again attempts a repair, until he realizes there is no one onthe other end of the phone.

In a subsequent interview the employee in Convention Services said "therehasn't been a call from Housekeeping all morning." He believed he was talkingto a guest in the transcript above. The novice speaker was too embarrassed toallow the tape to be played for other employees, so no data was collected aboutthis transcript in the lunchroom.

In Example (8) we again see the novice speaker trying to identify herself inT2. She uses the departmental name, rather than the person form"housekeeper"—an extremely common mistake among LPES housekeepers whowere frequently referred to with this form in phrases like "go tell housekeeping"or "get housekeeping." The "No" response in T3 suggests that the receiver has

6 See Schegloff and Sacks (1973) for more information about closings and presequences.7 A reviewer suggests that the use of "Good-bye" by the receiver also provides internal evidence thatthe receiver believes he is speaking with a guest, since in the example with the expert Englishspeaker supervisor there is no overt closing sequence at all and this is true. However, I am notconfident about the role of various types of closings. In Example (10), which is successful, there isalso a closing sequence after a negotiation of meaning. It seems possible that longer conversationduration or negotiation of meaning may also trigger the use of a closing. The same reviewer alsopoints out that, contradictorily, if the receiver believes he is talking to a guest he cannot, no matterhow tortured the call has become, hang up on a guest. I suspect that if we look at interactionsbetween native and normative speakers, we will find that such politeness gestures are frequentlyviolated, either by sales clerks who change their attention from normative speakers to nativespeakers, leaving the normative speaker unattended, to people who hang up when an impasse isreached. One particular testament to this as a strategy for ending difficult meaning negotiations wasillustrated a couple of years ago on Jay Leno when he telephoned random numbers to survey people.After reaching one normative speaker whom he was completely unable to communicate with, heabruptly hung up. This action didn't seem to strike the audience as odd, though he didn't hang up onanyone else. To me this suggests that hanging up may be seen as an acceptable strategy for ending adifficult meaning negotiation. Certainly, this is something to investigate more carefully.


interpreted this as a ((Claim not to have recognized the receiver)) rather than asa ((Claim that the receiver can recognize the caller)). Thus, we see anidentification attempt in T2. We also notice that the caller never gets tointroduce her topic of conversation before the call ends in T8.

(8) Second novice speaker: Unsuccessful call-in

TO: C: (Ring)Tl: R: Hello. Convention Services. How can I help you?T2: C: This is housekeepingT3: R: No. This is Convention Services. Shall I transfer you.T4: C: (.)T5: R: Let me transfer you?T6: C: (mumbled) housekeeping.T7: R: Uh. Let me transfer you.T8: (Ring, Ring)T9: C: (Hangup)

(Transcript of T. recorded in April, 1998)

In T2 of Example (9), the novice speaker tries to identify herself. In T4 shetries to introduce the topic, and is hindered by her inability to keep the floor. InT5 the receiver demonstrates an understanding of the request which is differentfrom the intended request. In T6 the LPES housekeeper is unable to change whathas been understood.

(9) Third novice speaker: Unsuccessful call-in

TO: C: (Ring)Tl: R: Convention Services. How can I help you?T2: C: This is L.

(•)Room 717(-)

T3: R: How can I help you?T4: C: need

(•)refrigerator(•)[pick uh-]

T5: R: [You— ] You need a refrigerator?T6: C: No. pee cup.T7: R: Ahhh,what?T8: C: Ping up'.

C).refrigerator. Ping up'.

T9: R: Hhhh. Can you say that again please?

190 Gibbs

T10: C: Ping cup.(-)717.(Tries to hand phone to the teacher who refuses it.)

Til: R: Uhh. 717.T12: C: Wait, please.

(Tries to hand phone to the teacher again who refuses it.)Thank you. Good-bye.(Hangs up)

T13: R: ((Ring)) ((Ring)) ((Ring))T14: T: (Teacher answers the phone after a short negotiation with the student and

clarifies the request.)

(Transcript of L. recorded in April, 1998)

In T2 of Example (10)', this novice speaker identifies herself ashousekeeping. It is nearly idiomatic usage of this term to refer to herself, excepther intonation which may have been a little bit off. The receiver begins torespond, although there is no way to know what he was going to say. Shemaintains the floor, however, and introduces her topic before ending her turn.

(10) Fourth novice speaker: Successful call-in

TO: C: (Ring)Tl: R: Convention Services. How can I help you?T2: C: Housekeeping. Pl[ease] pick-up roll-away.Tx: R: [Thi-]T3: R: Pick up roll-away?T4: C: Yes.T5: R: Which room?T6: C: 906.T7: R: 906?T8: C: Ok.T9: R: Ok. Bye. (Hangs up)T10:C: B-ai.

(Transcript of N. recorded in April, 1998)

In Example (11), the novice speaker identifies herself in T2 and is personal-ly identified in T3 by the receiver who tells her where her husband is. In T4 shesuccessfully introduces her topic despite a word choice error.

(11) Fifth novice speaker: Successful call-in

TO: C: (Ring)Tl: R: Hello. Convention Services. How can I help you?T2: C: Is S.T3: R: Hi. R is in the ballroom.T4: C: Ee. No. I need roll-away take out.T5: R: Hh. Is it a check-out?


T6: C: Si. Check-out.T7: R: Room number?T8: C: One One Zero One.T9: R: Alright.

(•)See ya.

T10:C: Bye. ((Hangs up.))

(Transcript of R. recorded in April, 1998)

Recording of the Novice Call-in After Modeling

The following model dialogue was created based on the expert speaker'stranscript in Example (6).

(12) Dialogue based on the expert speaker's transcript

TO: C: RingsTl: R: Hello.T2: C: Pick up in room , .

number item nameT3: R: .

numberT4: C: Yep. (Hangup)

Using a sheet of paper with the dialogue in Example (12), the same novicespeaker who was unsuccessful in Example (7) called Convention Services andproduced the transcript in Example (13), and the same unsuccessful speaker seenin Example (9) produced the transcript in Example (14).

The action sequences in (13) and (14) are nearly identical to the actionsequences produced in Example (6) by the expert speaker. There are only twodifferences between Example (6) and Example (13). First there is a request forthe topic in Tl by the receiver, and there is an appeal to politeness in T5 by thereceiver.

(13) First novice speaker: successful

TO: ((Rings))Tl: R:

T2: C:

T3: R:

Hello.Convention Services.How can I help you?Pick up refrigerator,Room 1215.

Room 1250?

((Summons))((Answer))((Display for recognition))((Request for intro. of topic))((Claim C has recognized R))((Claim R can recognize C))((Answer to request for topic))((Granting of request/

Uptake of topic))((Claim R has recognized C))((Question))

192 Gibbs

T4: C: Yes. Room 1215.T5: R: Thank you.

Good-bye. (Hangs up)T6: C: Good-bye, (laughter)

(J., May 1998)

((Answer))((Appeal to politeness))((Closure))((Closure))

In Example (14) the action sequences are again very similar to the actionsequences in Example (6). Despite severe pronunciation difficulties in T2, thereceiver is able to understand that it is an employee calling in and he respondsappropriately. The only real difficulty seems to be in T4 when the novicespeaker repeats the room number as a means of agreement instead of saying"yes" or "okay." This seems to confuse the receiver a little bit who then repeatshis request for confirmation of the room number.

(14) Third novice speaker: Successful

TO:Tl:

T2:

T3:

T4:T5:T6:

T7:

(L.,

((Rings))R:

C:

R:

C:R:C:

R:

May

Hello.Convention Services.How can I help you?Ping up. Room 709.

709?

709.709?Yes.Good-bye.Yeah.

1998)

((Summons))((Answer))((Display for recognition))((Request for introduction of topic))((Claim C has recognized R))((Claim R can recognize C))((Answer to request for topic))((Granting of request/Uptake of topic))((Claim R has recognized C))((Request for Confirmation))((Answer))((Repetition of request for confirmation))((Answer))((Closure))((Closure))

The unsuccessful recordings are all characterized by frustrating exchangesin which the LPES housekeeper loses the floor and the Convention Servicesemployee misidentifies the caller as a guest which the LPES housekeeper isunsuccessful in correcting due to an inability to regain control of theconversation and/or due to pronunciation or grammatical errors. The successfulrecordings also have pronunciation and grammatical errors in them, but theparticipants are all successfully identified as employees at the end of T2 andthey are able to coordinate a mutual goal.

Comparison of the Expert Speaker's Call-in and Canonical Structure

A comparison of the structure of the expert speaker's call-in in Example (6) withthe canonical structure of a social phone call (Schegloff, 1986), illustrated


Table 7.1. Action sequences of a call-in compared to the canonical structure

Expert Call-in Structure

TO: C: 1st: SummonsRing

Tl: R: 2nd:1st:

T2: C: 2nd:1st:1st:

T3: R: 2nd:2nd:

—end of call—

AnswerClaim that the caller can

recognize the receiverHello

Claim C recognized R.Claim R can recognize C

Introduction of TopicPick up in room 936,refrigerator.

Claim R recognized CUp-take of Topic

936

Canonical Conversational Structure


Tl: R: 2nd: Answer1 st : Claim that the caller can recognize

the receiverHello

T2: C: 2nd: Claim C recognized R1 st: Claim R can recognize C1st: Greeting

Hi.

T3: R: 2nd: Claim R recognized C2nd: Greeting

Oh hi::

T4: C: 1st: Introduction of topic

T5: R: 2nd: Up-take of topic

in Table 7.1, reveals three main differences. First, in the expert's call-in, thetopic is introduced preemptively at T2 before the identification sequence is evencompleted. Second, there is no personal identification of the caller and receiverand the identification sequence is a nonovert sequence. Third, there is nogreeting sequence in the hotel data.

Although there is a hello in Tl of Example (6), it is not considered agreeting for reasons that Schegloff (1986) takes up in detail. Briefly, it isconsidered an answer to a summons. (That is why, according to Schegloff, thereare usually three hellos at the beginning of a phone conversation. The first hellois an answer to a summons and not a greeting. The second and third hello are thegreetings.) In this institutional data, in the expert speaker's transcript, there isnothing that is construable as a greeting.

Inasmuch as I am claiming that there is not a greeting sequence in the hoteldata, based on the omission of any overt statements that seem like true greetings,one might wonder whether there is also really an identification sequence. Theexistence of the identification sequence is primarily based on the same logicalargument provided before by Schegloff, but with one modification: In the hoteldata, people are not being identified as particular individuals. Rather, they arebeing identified as belonging to a particular group (such as Housekeepers,Convention Service Employees, Room Service, or Guests). The exception to thisis Example (11) in which R is recognized as an individual, but the response tothis is fascinating. The receiver immediately tells R where her husband is,

194 Gibbs

suggesting that he has oriented to the caller as an outsider—a colleague's wife—not as a colleague herself seeking to coordinate collegial duties.

The fact that group identification has occurred in the collegial exchangeswas also confirmed by the interviews after the phone conversations. After eachphone conversation, the participants had a clear understanding as to whether ornot they had been talking to an employee or to a guest. It should be noted,however, that in the collegial exchanges participants generally did not seem tohave a name or face attached to the person they were talking to.

Speech Events Compared:The Expert Speaker vs. the Novice vs. Schegloff

Table 7.2 allows us to compare the structure of the novice speaker's unsuccess-ful call-in from Example (7) with the structure of the expert speaker's call-infrom Example (6) with the canonical structure. Like the expert speaker's call-in,the novice speaker omits the greeting sequence. Unlike the expert speaker's call-in, the novice speaker follows the canonical structure in having an overtidentification sequence and identification of person. Finally, and perhapscrucially, while the expert speaker's call-in introduces the topic preemptively inT2 before the identification sequence is complete, the novice speaker does notintroduce the topic until T6, after trying to complete an identification sequence,just as in the canonical telephone call.

Misidentification of the novice speaker seems to occur when T2 passeswithout introduction of the topic and without the Convention Services employeeorienting to either the housekeeper or the need to coordinate a task. In additionto T2 passing without introduction of the topic by the novice, the personalidentification sequence offered by the novice in T2 corresponds to the canonicaltelephone call examined by Schegloff, but not to the call-in genre. While anovice native speaker might have been able to repair this misidentification, thisLPES novice speaker was unsuccessful in his repair attempts.8

Effect of an Accurate Model

Providing the novice speaker with an accurate model did in fact result in asuccessful call-in, despite the persistence of marked phonological difficulties,since the phone call resulted in an employee from convention services coming tothe requisite floor where the housekeeper directed him to the correct room. Thesuccess of the novice speakers when using the model produced by the expertspeaker adds validity to the idea that genres do have unique rules that areutilized by expert members of the discourse community to accomplish theirgoals, rules that can be taught to novices.

During the first recording, the students in this study were using a structurethat included an overt identification sequence and they were not immediately

8 For more about why he was unsuccessful in repairing the misidentification, see Gibbs (2002).


Table 7.2. Expert and novice hotel calls compared to a social call.

Hotel data: Expert supervisor


T1:R: 2nd: Answer1st : Claim that the caller canrecognize the receiverHello

T2: C: 2nd: Claim that C hasrecognized R1st : Claim R can recognize C1st : Introduction of topicPick up in room 936,refrigerator.

T3: R: 2nd: Claim that R hasrecognized C2nd: Up-take of topic, grantingof request936

--end of call--

Speech acts which areperceived differently by thecaller and the receiver areindicated with bold and italicsrespectively.

Hotel data: LPES speaker


T1:R: 2nd: Answer1st : Claim that the caller canrecognize the receiverHello. Convention Services.

T2: C: 2nd: Claim that C hasrecognized R1st: Claim R can recognize C

This is J. I'm- (.8)[housekeeping]

T3: R: 2nd: Claim that R hasrecognized C1st: Request for topic[How can-]How can I help you?

T4:C: 1st: Repair of T3R: 2nd :Intro. of topic: answer1st:Request for housekeeping

I am housekeeping.

T5 :R: dispreferred 2nd: Uptakeof topic.1st: DirectiveC: dispreferred 2nd : ananswer to the repair whichindicates that the repair wasunsuccessfulTo call housekeeping youneed to dial 52.

T6: C: 1st: Repair of T3:Claim that R has recognizedC: assertion: Claim that Rcan recognize C1st: Introduction of topic:request for pick-up2nd: Rejection of directive tocall housekeepingIst: Request for housekeepingHousekeeping.Pick up.

Social call: Schegloff data


Tl: R: 2nd: Answer1st : Claim that the caller canrecognize the receiverHello.

T2: C: 2nd: Claim that C hasrecognized R1st: Claim R can recognize C1st : GreetingHi.

T3: R: 2nd: Claim that R hasrecognized C

2nd : Greeting

Oh hi::

T4: R: 2nd: Answer1st: How are you

Fine. And you?

T5: C: 2nd: Answer1- : Intro. of topic

Fine. Say, are you busy later?

T6: R: 2nd: Up-take of topicNo. What's up?

196 Gibbs

introducing the topic. Perhaps this was because an overt sequence is required insocial phone calls and topics are not introduced until after identification hasoccurred. It also could have been an effect of prior training on incorrect models.Both the dialogue practiced in the previous ESL class, Example (15), and theNew Employee Orientation video, Example (16), included overt identificationsequences which the expert speaker did not use, and both waited to introduce thetopic until after identification had occurred. This suggests that it is important tobase instructional materials on talk produced by expert members of thediscourse community so that the rules of the genre can be incorporated in thetraining model.

(15) Dialogue from previous course materials

C: ((Ring))R: Hello. Conference Services, this is Jeff.C: Hi. This is Mary.R: How can I help you?C: I am in room 100. Can you come get a refrigerator, please?R: Sure.C: Great. Good-bye.R: Good-bye.

(16) In-house, new employee orientation video script

TO: C: ((Ring))Tl: R: Hello, Conference Services.T2: C: Hello, this is Mary from Housekeeping. I'm in room 214. There is a

wet bar and a refrigerator here which need to be picked up.T3: R: Thank you, Mary. Do you need the room right away?T4: C: Yes, I do.T5: R: Ok. We'll send someone up. Can I do anything else for you?T6: C: No. Thank you. Good-bye.T7: R: Good-bye.

(Training video for new housekeepers, circa 1997)

New employees usually learn the genre by following around a seniorhousekeeper for 3 days, listening to their mentor perform call-ins, and practicingcall-ins themselves. The LPES housekeepers are usually trained by a nativespeaker of their language (so new Vietnamese housekeepers are trained bysenior Vietnamese housekeepers). Because most of the nonnative speakers ofEnglish were not successful in making call-ins, new LPES housekeepers lackedgood mentorship for learning the call-in, although they probably did hear theirsupervisors making call-ins. As the LPES housekeepers may have beeninfluenced by the incorrect models displayed during the new employeeorientation and in a prior ESL program, it is not clear whether they would havelearned the genre on their own if they hadn't had bad models in theirenvironment or whether this genre was inaccessible for other reasons.


Comparison of the Call-in with the Doctor's Office Calland the 911 Call

Table 7.3 compares the sequences found in the hotel data with the sequences ininstitutional phone calls studied by Hopper et al. (1990) and Wakin andZimmerman (1999).

All of the institutional samples show preemptive introduction of topic, sothis may be a feature of institutional telephone calls. Overt vs. nonovertidentification seems to vary by institution and may be specific to the individualgenres. In all of the transcripts of telephone calls to a work place shown here,the greeting sequence is omitted. The idea that the greeting sequence may not bemandatory actually receives some support from Schegloff (1986). Schegloffnotes that he finds the job of the greeting sequence difficult to define.9 However,he suggests, it seems to be a ritual whose purpose is orienting the caller andreceiver to one another and it is used to establish cooperation toward a sharedgoal. Since genres have their own sets of rules, it seems logical that an orienta-tion to the genre must occur. Thus, it might be more appropriate to say that at T2an orienting process occurs. In the social, hello-hello sequenced phone calls thisorienting process between individuals is realized by a "greeting" sequence. Inthe case of the call-in, however, this orienting may have to do with displayingthemselves as employees making a call-in by introducing the topic preemptivelywith a directive.

The how-are-you sequence is also not in any of the transcripts recordinginstitutional interactions. In fact, it seems likely that the how-are-you sequenceis only tangentially part of the opening sequence even in social calls, in that itseems to be some sort of presequence to the introduction of the topic in social

Table 7.3. Comparison of Various Institutional Telephone Calls

Feature

Summons-answer

Identificationsequence

OvertNonovert

Greeting

How are you?

Hotel employee-expert

Yes

Yes

NoYes

No

No

Hotel employee -novice (Unsuccessful)

Yes

Yes

YesNo

No

No

Client-to-doctor

Yes

Yes

YesNo

No

No

Client-to-911

Yes

Yes

NoYes

No

No

9 This is an oversimplification of Schegloff's analysis. He writes "the jobs attributed to greetingsdefy listing, let alone description, here; at a minimum, they put the parties into what Goffman (1963,p. 100) has called a ritual state of ratified mutual participation, and in doing so may accomplish otherwork for the interaction and its parties as well" (Schegloff, 1986, p. 118).

198 Gibbs

phone calls displaying intimacy. Hopper observes that intimate social callscontain substantially more presequences than phone calls between nonintimates.Thus, the institutional phone calls seem to display canonically a summons-answer sequence and an identification sequence with introduction of the topicprior to the completion of the identification sequence. The orientation sequenceseems to be achieved in unique ways by different genres. Different genres alsomay differ in regards to whether the identification sequence is handled overtlywith its own turn constructional unit, or nonovertly.

CONCLUSION

These data are in agreement with the notion that genres have their own rules for(1) which actions are engaged in and (2) the sequence in which the actions areengaged in, and that training materials need to be designed with this awarenessin mind since not following the rules of the genre can result in unsuccessfulspeech events. Unlike Schegloff's social telephone calls where preemptiveintroduction of topic before completion of the identification sequence mightindicate irritation, preemptive introduction of topic in the call-in genre seems toindicate that the caller is a fellow employee, rather than a guest. Failure tounderstand and follow the rules of the genre seems to result in an unintendedmeaning being transmitted: specifically "I am a guest" rather than "I am a fellowemployee." Due to the difficulty experienced by the LPES housekeepers incontrolling the talk, such a misidentification became difficult for them to repair.

This work suggests that there are phone call genres that have featuresdifferent from the call features described in Levinson (1983) and Schegloff(1968, 1979, 1986). In particular, work needs to proceed on the differencesbetween employees and clients when they call a business. Recordings of newemployees acquiring discourse competence in the rules of their new speechcommunity may also yield interesting insights on acquisition and proto-typicality. Finally, the generality of the preemptive introduction of topic ininstitutional data, the meaning of overt-nonovert identification sequences ininstitutional data, and processes for orienting to the type of call needs to beestablished across genres.

REFERENCES

Atkinson, J. M., & Heritage, J. (Eds.)- (1984). Structures of social action. Cambridge, England:Cambridge University Press.

Gibbs, T. L. (2002). Misidentification of limited proficiency English speaking (LPES) employees inhotel call-ins. Unpublished Plan B paper, University of Minnesota.

Goffman, E. (1963). Behavior in public places. New York: Free Press.Fairchild, M. (1995). Mind your modals: The telephone request in a business context. Unpublished

Plan B paper, University of Minnesota.Hopper, R., Doany, N., Johnson, M., & Drummond, K. (1990). Universals and particulars in

telephone openings. Research on Language and Social Interaction, 24, 369-387.Levinson, S. (1983). Pragmatics. Cambridge, England: Cambridge University Press.Markee, N. (2000). Conversation analysis. Mahwah, NJ: Lawrence Erlbaum Associates.


Psathas, G. (1995). Conversation analysis. Thousand Oaks, CA: Sage.Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organization of

turn taking in conversation. Language, 50, 696-735.Schegloff, E. (1968). Sequencing in conversational openings. American anthropologist, 70, 1075-

1095.Schegloff, E. (1979). Identification and recognition in telephone conversation openings. Everyday

language: Studies in ethnomethodology (pp. 23-78). New York: Irvington.Schegloff, E. (1986). The routine as achievement. Human Studies, 9, 111-151.Schegloff, E. (1992). On talk and its institutional occasions. In P. Drew & J. Heritage (Eds.), Talk at

work: Interaction in institutional settings (pp. 101-134). Cambridge, England: CambridgeUniversity Press.

Schegloff, E. (1996). Turn organization: One intersection of grammar and interaction. In E. Ochs, E.Schegloff, & S. Thompson (Eds.), Interaction and grammar (pp. 52-125). New York:Cambridge University Press.

Schegloff, E., & Sacks, H. (1973). Opening up closings. Semiotica, 7, 289-327.Swales, J. (1990). Genre analysis. Cambridge, England: Cambridge University Press.Wakin, M., & Zimmerman, D. (1999). Reduction and specialization in emergency and directory

assistance calls. Research on Language and Social Interaction, 32, 409-437.Whalen, M., & Zimmerman, D. (1987). Sequential and institutional contexts in calls for help. Social

Psychology Quarterly, 50, 172-85.Whalen, M., & Zimmerman, D. (1990). Describing trouble: Practical epistemology in citizen calls to

the police. Language in Society, 19, 465-492.Zimmerman, D. (1992). The interactional organization of calls for emergency assistance. In P. Drew

& J. Heritage (Eds.), Talk at work: Interaction in institutional settings (pp. 418-469).Cambridge, England: Cambridge University Press.

8

Practical Considerations

Kathleen Bardovi-HarligBeverly S. HartfordIndiana University

This chapter discusses the practical concerns of getting started in research thatexplores interlanguage pragmatics through institutional talk. Rather thandrawing solely on our own experiences, we have asked colleagues to share theirexperiences with us, and we report these in this chapter. Some have conductedresearch in the institutions represented in this volume, but some have done theirwork in other institutions as well, allowing for a broad perspective.1

The institutions represented in this chapter include universities—specifically, the writing center (Williams, chapter 2, this volume), classrooms(Davies & Tyler, chapter 5, this volume), university advising sessions (Bardovi-Harlig & Hartford, 1990, 1993, 1996; Hartford & Bardovi-Harlig, 1992)—secondary schools (Yates, chapter 3, this volume), employment agencies(Kerekes, chapter 4, this volume), a four-star conference hotel (Gibbs, chapter 7,this volume), the INS (the United States Immigration and NaturalizationService; Johnston, 2003a, 2003b, in press; Seig & Winn, 2003; Winn, 2000,under review), hospitals and clinics (Cameron, 1998; Cameron & Williams,1997), and a variety of workplaces (Clyne, 1994). Most of the research settingswere in the United States, but information from Yates and Clyne suggests thatthe process of setting up research on institutional talk is similar in Australia. Theguidelines presented here are intended to help researchers get started; weacknowledge that specific requirements may vary by institution and localpractices. Comprehensive reports such as Clyne (1994) contain detailedinformation and should also be consulted. We have organized this chapter by theapproximate sequential order of steps that one may take in organizing such aproject. These include identifying the institution, getting permission from theinstitution, completing a review process, collecting data, and winding up theproject. Necessarily there is some overlap among the sections which reflectsoverlapping considerations at different points in a project.

1 We thank the following colleagues for their expertise: Tara Gibbs, Julie Kerekes, AndreaTyler, Jessica Williams, Lynda Yates, Michael Clyne, Richard Cameron, Alexandra Johnston, MaryTheresa Seig, and Michelle Winn.

In the remainder of this chapter, any unattributed quotation should be understood as a responseto a set of questions sent out through e-mail in July and August, 2003. These questions were sent tothe authors whose work is included in this volume as well as to other researchers who have studiedinstitutional talk of L2 speakers.

201


IDENTIFYING THE INSTITUTION

There seem to be two main orientations to studying institutional talk. In one,researchers study their local institutions, making an investigation of theinteractions around them. In the other, investigators seek out new institutionswith which they have no official relationship.

Guideline 1. Make your institution work for you. Look at yourlocal setting as an outsider. Evaluate the interaction patterns.What are the services offered by your unit and who uses them?What type of talk occurs there?

Our own research on the academic advising session has been of the firsttype, a study of the local institution. The academic advising sessions that westudied were meetings between faculty members serving as academic advisorsfor graduate and undergraduate students in the department of linguistics(Bardovi-Harlig & Hartford, 1990, 1993, 1996; Hartford & Bardovi-Harlig,1992). We began work on the academic advising session because we wereintrigued by the different outcomes of the sessions and were interested to learnhow talk during the sessions influenced or determined those outcomes. Naturallocal settings for academic researchers who teach at universities are often (butnot always) the universities themselves and this means studying institutional talkat the university. It is important to keep in mind, however, that university talk—like the university itself—is made up of talk in many different settings.Universities offer many services and, consequently, offer many differentopportunities to investigate talk in different local settings including somedescribed in the present volume such as the university writing center and labsessions. In addition to these and the faculty advising sessions already discussed,other university settings include peer advising (He, 1994), class lectures (Tyler,1992), the local Intensive English Program (Bardovi-Harlig & Hartford, 1997;Hartford & Bardovi-Harlig, 1996b) and office hours (Tyler, 1995) Not allinsider research is done at universities, however: In this volume, Gibbs reportson data that she collected when she was teaching ESL at a four-star conferencehotel.

Guideline 2. Consider a range of institutions.

Research done outside researchers' own institutions shows a remarkablearray of settings. We were intrigued to find not one, but two studies presented on

Practical Considerations 203

talk during INS interviews at the 2003 meeting of American Association ofApplied Linguistics (AAAL). These studies took place in different states(Johnston, 2003a, 2003b, in press; Seig &Winn, 2003; Winn, 2000). Otherstudies involving native and normative speakers were set in a psychiatric ward ofa major metropolitan hospital (Cameron & Williams, 1997), a variety offactories (Clyne, 1994), a social services office (Tarone & Kuehn, 2000), andsecondary-school classrooms (Yates, chapter 3, this volume). Readers areencouraged to consult Clyne (1994) for a detailed discussion of selecting worksites and participants.

Guideline 3. Make use of your social network.

Conducting research away from one's home institution requires makingcontacts. This ranges from relying on one's network of friends to help makecontacts to making "cold calls," that is, approaching an institution without aprior personal relationship. Many researchers report that friends helped themmake contacts within the targeted institution. Yates, who studies the talk ofteachers in secondary classrooms, was involved in teacher training at theuniversity level (this volume, Chapter 3). Her department sent some of itsstudents to some of the schools that she eventually studied. She explains, "Ideveloped a relationship with people from other universities who werecoordinating practicum placements in school. Through these I identifiedparticipants. When they were assigned to schools for their practicumplacements, I contacted the principals of the schools and the relevant teachers torequest their permission." Like Yates, Cameron, whose research followed nursesat four different hospitals and clinics, developed his contacts through facultycolleagues. Cameron reports, "I found out names of contacts through differentfaculty members in the School of Nursing at Penn. Then, I called on the phoneand went in person to explain the project."

Kerekes' contact in the employment agency industry began with a friendwhose sister lived in another state. Kerekes recalls, "I had a friend whose sisterworked in the employment agency industry as a staffing supervisor in Seattle.She and I communicated by e-mail and phone and she expressed interest in mystudy, so she connected me with some of her colleagues in a San Francisco BayArea branch of the same employment agency for which she worked."

Guideline 4. Treat the institutional representatives as experts.

Treating the institutional representatives as the experts regarding theinstitution and not as hindrances to the study helps in obtaining permission. It is


important to keep in mind that the researchers may have much to learn from theinstitutional representatives themselves (cf. Douglas and Selinker's, 1994, use ofthe subject specialist informant).

Guideline 5. Once you have a contact at the institution, learnthe chain of command, and then follow it.

Making contacts and checking with supervisors is important whetherworking in your own institution or an outside institution. Don't assume you willautomatically be granted access to interactions either in your immediatedepartment or affiliated departments. Take the time to make the right contactseven at home; securing permission at all steps along the way helps the researchproceed more smoothly.

Interestingly, the research chain of contact often seems to start with anindividual at the institution, then move up to a supervisory or higher institutionallevel (or levels), and then return to the level of the client. Cameron explains thathis approach to clinics and schools is similar: "I arranged it first with a localadministrator and then with the specific individuals in the clinic where I wasgoing to do the research. I have done similar work in schools. First, permissionfrom the principal and school district. Then, permission from individualteachers. Then, permission from parents and kids. But, where possible, I tried tomake a local contact first, then sought out the higher up administrator. In myexperience, if I can go to a higher up administrator, saying that I had spokenwith a nurse (who had approved the idea), I seemed to have a smoother ride inobtaining permission."

Clyne conducted his research on workplace talk in eight different settings, acar factory of American origin, a car factory of Japanese origin, a textile factory,an electronics factory, a catering department, an education office of agovernment operation, an employment office, and a meeting at a multiculturalparents' group at a high school. Clyne reports that they "discussed the project,what was required, and our responsibilities (anonymity, confidentiality,opportunity to withdraw, possibility of receiving information on the findings)with management, union representatives, shop stewards, and workersthemselves."

Yates presents the order of contacts that she used when researchingsecondary schools in her position as a researcher at a university, "I had to getethics clearance from my university (filling out appropriate form), thenpermission from the regional director (by letter with accompanying outline ofthe project), then from the principal (letter plus details), then the supervisingteacher (phone) and then from the parents of all the students in each class(permission slip to be given out to students in advance and returned by the dateof the taping)."


Guideline 6. Be persistent and patient.

Johnston's account of her contacts with the INS illustrates several points:that research can be conducted in a full range of settings, that researchers cancultivate a relationship through cold calls when necessary, and that the chain ofcommand is very important. Most of all, Johnston's account emphasizes thatpersistence and patience pay off. Johnston recounts, "I made 41 telephone callsto the Public Affairs Office, sent five letters describing my project to differentlevels of managers up to the District Director, and presented my project at threemeetings with INS personnel before I won permission to present my project tothe entire staff of District Adjudications Officers. The process took over sixmonths."

Guideline 7. Consult with other researchers, but don't let nay-sayers discourage you.

All researchers benefit from the expertise of colleagues, just as we havedone in this chapter! Colleagues who have done institutional research can offermany practical suggestions. However, as Johnston learned, some advice can beunnecessarily discouraging. She writes, "Don't get discouraged if otherresearchers tell you that a certain approach to gaining access won't work. Isought help from several researchers who had successfully gained access to theINS, which rarely grants permission for on-site research. All of those whom Icontacted had gotten in through connections or a friend-of-a-friend. They toldme this was the best way and that cold calling the Public Affairs Office wouldn'twork because it's the job of Public Affairs to screen people out (they are thegatekeepers to the gatekeepers). That may be true, but I think persistence,courteousness, and a willingness to provide a service in exchange for accessgoes a long way. It worked in my case."

GETTING PERMISSION TO CONDUCT THE STUDY

Once contact is made, the next step is to convince the institution to givepermission to conduct the study. We asked what the researchers felt were theconvincing factors for their gaining permission to do their studies. The mostimportant, reported by all, was that they stressed the benefits of their project tothe institution. In many cases, there were direct benefits, such as designing atraining course for institutional members, or materials development for thosemembers to use. In others, offering to share the results with the institution and


pointing out how the results might be used to help institutional members wereconvincing factors. In all cases, some offer of help to the institution as a result ofthe study was felt to be a major factor in obtaining permission.

Guideline 8. Cite the benefits to the institution.

Yates explained to the school board and the candidate schools the benefit oflearning more about the language needs of teachers. Such knowledge has thepotential of improving teacher education to the potential benefit of both schoolsand teachers. A second type of benefit accrues directly to the participatinginstitution rather than to the field or profession more generally. Like Johnston,Winn also studied immigration interviews at the INS. They took slightlydifferent approaches. Johnston convinced her point of contact that her primarygoals were to help the INS and to complete her degree. Winn involved theinstitution in the planning process by asking them what they would like to learnfrom the research, "I wrote a letter outlining my reasons to study thenaturalization interview and asked if the INS wanted me to focus on anyquestion in particular in my research. I explained the personal stake I had in theresearch—having been a U.S. citizenship preparation teacher."

A different type of benefit to the institution is the researcher's offer to helpout in some way. In this way, the researcher also learns more about how theinstitution works. Kerekes, for example, volunteered at the employment agencywhere she collected her data. She helped out at the front desk, helped with filingin the back room, and answered phones. She reports that she offered to do somevolunteer work in order to get a partial insider's perspective while collectingdata.

Guideline 9. Be personally invested.

Winn, in her discussion of working with the INS, suggests that researchers"be personally invested." As Winn's advice suggests, it is also helpful to let theinstitution know that you are invested in the field. Teacher-researchers oftenhave an obvious link to various institutions. Winn's link to the INS as acitizenship teacher might have made her interest in the INS more tangible to theapproving officer. Similarly, as a teacher educator, Yates also had a demonstra-ble interest in the language that teachers need, as Cameron did when he wasdeveloping courses for nursing students who were normative speakers ofEnglish.


Guideline 10. Do not hinder the everyday workplace activities.

The researchers noted that assurances of unobtrusiveness on their part whileconducting the study were also important. Institutions have goals to accomplishand work days which are allocated to a variety of activities, and they are noteager to disrupt the flow of work. Thus, unobtrusiveness in the form of notaffecting the organizational practices in any major way was appreciated. Peoplecould essentially carry on with what they would normally do without having tochange time-lines, spatial locations, and so forth. At the same time, an actuallow level of intrusion on the part of the investigator seemed desired. Setting uprecording equipment ahead of time, locating oneself in a fairly unobtrusive partof the room (or not being present at all), not participating in the process at hand,were assurances of noninterference that various researchers gave which they felthelped to persuade the institutions.

Sometimes the agreement for unobtrusiveness is general as Kerekes reports,"My presence was welcome under the condition that my work would in no wayslow down or impede on FastEmp's normal procedures regarding the selectionof job candidates." In other cases, unobtrusiveness relates to the recordingequipment that researchers use. Cameron could only use audio recorders,without attaching microphones directly to the patients. The student-teachers inYates' study wore portable recorders that she describes as a "walk person" plusmini mike for each class recorded, and she worked the video camera from theback of the room. When we recorded the advising sessions we quickly learnedthat both advisors and students preferred a small tape recorder with an omni-directional microphone to a larger, higher quality stereo recorder with separatelapel mikes for each participant. (We return to the topic of equipment later inthis chapter; see Guideline 19). Setting up equipment in advance of theinstitutional event is often part of being unobtrusive. This minimizes thedisruption to the normal flow of the event.

Guideline 11. Be accessible.

Make your proposal accessible to the institution. Researchers may oftenforget that others do not necessarily share our interest, expertise, or vocabularyfor talking about language. Bringing the institution into the research as a partneris an important step. As Johnston suggests, "translate your research agenda intoaccessible language that dovetails with the internal concerns of your targetinstitution. Show them that you understand (or want to learn about) their


concerns and questions. Offer your problem-solving services, withoutcompromising your goals or ethics."

Guideline 12. Guarantee and deliver fair treatment.

Some institutions may be particularly sensitive to public opinion. Such maybe the case at schools which have recently been in the news, medicalinstitutions, and government agencies. For example, Johnston reports that "theINS has received a great deal of criticism from government and the media for itsinefficiency and errors. Representatives of the institution are understandablywary of outsiders who request to research and write about its practices. Thiswariness was apparent in my initial contacts with Public Affairs Officer,Management, and District Adjudications Officers. Therefore, I was careful topresent my research as an instrument that might provide benefit to the INS interms of improving service to clients and reducing miscommunication withclients. I offered to use my results to create a training module for the officers.And I reassured them that I was not affiliated with the press or any associationadvocating immigration reform (a question I was asked more than once). Afterreading my confidentiality forms and engaging in multiple interviews, I finallyconvinced my point of contact that my primary goal was (1) to help the INS and(2) to complete my degree (and not to excoriate the INS in my futurepublications, which I intended for an academic audience, not the mass media)."

Guideline 13. Contact the institutional representatives.

Once the institution has approved the study, then permission remains to besecured from two more groups, the institutional representatives and the clients.These are the people whose interactions will be recorded. Generally the first stepis to contact the institutional representatives. Sometimes these are the sameindividuals that were contacted informally at the outset of the study, but this stepsecures their official permission.

In our observations of the academic advising sessions, we presented ourcase at a faculty meeting, because in our department, all faculty membersparticipated equally in the advising of graduate and undergraduate students. Werecruited five faculty participants in addition to ourselves. As experiencedadvisors, we participated in the first year of advising prior to data analysis. Oncewe began to analyze the data, we no longer took part in the tape-recordedinterviews. The five volunteers constituted more than half of the faculty at thetime.


Williams had the advantage of being the faculty director of the writingcenter in which she conducted the research, although she reports that the tutorswere quite eager to participate. (There was also an administrative director whowas interested in having research conducted in the center). They were recruitedby word-of-mouth from among the tutor-staff. There was no problem findingtutors who wished to participate. Williams is the only researcher we contactedwho reports being able to pay her tutors for participating. The Vice Chancellor'soffice had also funded the project. As a result, she had not only institutionalpermission, but institutional support, much as Gibbs had at the hotel.

While word of mouth recruiting worked among the writing center tutors,other institutional participants were recruited by face-to-face meeting or byphone calls. Kerekes met with staff personally: "After the proposal wasaccepted, I spoke with all of the FastEmp staff to ask them if they were willingto participate in my study, and then had them sign consent forms." Johnston alsomade a face-to-face appeal for participation at a staff meeting. She reports,"After receiving approval from management, I was allowed to present myproject at a staff meeting of the nine District Adjudications Officers in order toinvite their participation. One officer volunteered."

Yates contacted her participants by phone using contact details that wereprovided by the practicum placement coordinators in each university where theywere undertaking their teacher training. One of the factors that leads to differentapproaches is whether the research is located in a single or in multiple settings.Johnston's and Kerekes' studies took place in a single office of the INS andFastEmp, respectively, whereas Yates' prospective student-teacher participantsattended different universities and would be assigned to different secondaryschools.

Guideline 14. Contact clients of the participating institutionalrepresentatives.

The institutional representatives form the bulwark of the project. Once theinstitutional representatives have agreed to participate, then the task ofcontacting the institutional clients begins. As we discussed earlier in chapter 1,clients are identified through their interaction with the participating institutionalrepresentatives. Thus researchers variously approach students who come toparticipating advisors, instructors, or tutors, patients who seek the services ofparticipating nurses, or visa applicants who are assigned to a particular INSofficer. Most of the contacts with clients were done face-to-face in theinstitution, where potential participants could be asked if they would take part inthe study. This might be done in a one-on-one context, or in a more generalmeeting of possible participants where the researcher would present the project.If the parties agreed, they were usually given a consent form of some type to


sign. In those cases where the researcher was also a member of the institutionoften the potential participants were initially contacted through recruitingmeasures such as distribution of flyers and announcements, but were still askedto sign a consent form. In only one case discussed in this book (Williams,chapter 2, this volume), was payment offered for participation in the study.Payment did not seem to be a primary motivating factor in these studies,however.

The writers in Williams' writing center studies were one group that couldbe recruited by group contact. Williams recruited student participants (theinstitutional clients) through undergraduate composition classes by sendingflyers to the classes, after which her graduate assistant would "talk it up." Suchrecruitment procedures paralleled the general practices for encouraging studentsto use the writing center. In contrast, Kerekes, Cameron, and Johnston met theirprospective client-participants at the time of the institutional encounter andasked them each to participate face-to-face. Their approaches varied accordingto the institutions. Cameron used slightly different approaches in the hospitalsand clinics that he studied. At three of the sites Cameron reports, "I asked themface to face and had them sign short, small consent forms. For patients, in theHMO [Health Maintenance Organization] clinic, the nurse would first enter andask if I could speak to the patient about the project."

Kerekes outlines what approaching a prospective participant entails: "Icontacted job candidates individually, at the time that they had their FastEmpjob interviews. At that time, I introduced myself, described my project ingeneral terms ('I am a graduate student at Stanford, working with FastEmp tounderstand how the interviewing process can be improved'), explained that theirparticipation in my study was voluntary and would not affect their success orfailure, and then I asked them if they would like to participate."

Because applicants at the INS are routinely informed by INS that theirinterview will be taped, Johnston secured permission to use these tapedinterviews after the session (Johnston) When the officer led an applicant into theoffice for their interview, he turned on both the camcorder and audiocassetterecorder in view of the applicant (and their spouse, children or attorney). (AllINS officers have the prerogative to videotape and otherwise record interviews,and all applicants are apprised of this in the official INS letter that notifies themof their interview date and time in a sentence that reads: "This interview will bevideotaped.")

After the officer concluded the official interview, he asked the applicants ifthey would please step into the office next door, where I was waiting. (I wasgiven the use of an empty office to conduct my interviews and to staythroughout the workday.) He told the applicants that I would like to use thevideotape in my research, and that I would explain further. All applicants (60)obligingly stepped into my office for a 5 minute explanation. If they did notagree to participate, I erased their video and audiotapes.


If they agreed to participate, I explained the consent form and its contents(right to confidentiality and anonymity, right to withdraw, right to contact theresearcher at any point, etc.). After they read the form and signed it, I conducteda brief interview about their experience communicating with the INS officer andtheir INS experience in general. In all, 51 of 60 applicants agreed to participate."

Guideline 15. Note that post-interviews may be possible.

Kerekes (chapter 4, this volume) and Johnston (2003a, 2003b) not onlyrecorded the employment and INS interviews, respectively, but they alsoconducted postevent interviews with the participants, a step that involves aneven higher level of participation by both representatives and clients. Johnston'saccount of her procedure was reported in the immediately preceding section.

STANDARD REVIEW PROCESS

Universities have review processes that protects human subjects. At someinstitutions this is called the Human Subjects Committee, in others the internalreview board or IRB. The institution at which you plan to work may also haveits own version of such a process (this is true of public school districts, forexample). This should not discourage situated research. A researcher must alsoseek approval from the review board for any study in interlanguage pragmatics,including, for example, the completion of questionnaires.

You will need a consent form for all parties involved, specifically, theinstitutional representative and the client. Very often, IRB procedures specifythe format of a consent form and will provide a template for you to use. Thesecan be difficult for nonnative speakers to understand, and outside the university,such consent forms may be difficult for any participant to understand. To thatend, where permitted, Winn suggests writing the consent form in "PLAINENGLISH" (emphasis in original). Researchers whose institutions do not havesuch requirements should consider following similar general guidelines out ofcourtesy for the participants.

Included in considerations when working with human subjects are confiden-tiality and anonymity. It is standard to guarantee anonymity to both institutionalclients and representatives. This often involves using code names and destroyingidentity markers, not showing the raw data to anyone outside of theinvestigators, keeping the data locked away, among others. As an illustration,Johnston reports that "the applicant consent form and the officer consent formpromised both applicants and officers confidentiality and anonymity. Theyaffirmed that the results of the study were intended for academic publication andpresentation, and that no names, personally identifying characteristics or


locations would be revealed.... The tapes became the property of the researcherand were to be kept in a safe place with restricted access."

Johnston also reports creating a third type of confidentiality that extended tothe INS, allowing the institution the same rights as the individual participants:"to view the tapes and request partial or full erasure in the interest ofsafeguarding the confidentiality and anonymity of both the INS and the INSofficers involved. This form allowed an institutional representative (orrepresentatives) named by the INS to have access to the raw video, audio andwritten field data in which any INS officers appeared." The same approachcould be taken with other institutions. Institutions that have not previouslyparticipated in research may be wary of having interactions taped. Such ameasure might help reassure them.

Researchers also often create a pseudonym for the institution, providinganonymity for the institution as well as for the institutional representatives andclients. Kerekes named the employment agency "FastEmp," a pseudonym thatcaptured the type of institution and the potentially fast placements that theindustry and the clients sought. Clyne (1994) refers to his workplace locationsby generic types such as Catering, Weavers, Education Office, and EmploymantOffice. Cameron describes his research sites by their main characteristics: apsychiatric unit of a major metropolitan hospital, a community-based HMO(health maintenance organization), a neurotrauma unit of a major urban hospital,and a gynecological and birthing clinic of a smaller metropolitan hospital.

Internal review approval also typically requires that participants be able towithdraw from the study at any time. In the case of single institutionalinteractions (as opposed to longitudinal studies), we generally offer participantsthe option of turning off the recorder at any time during an interaction.Occasionally students took us up on this option during advising sessions orimmediately following the advising session proper in order to discuss sensitiveissues.

Participants may also request that their tape(s) be returned to them ordestroyed. (These are standard items on internal review board applications forapproval.) Although 51 of the 60 applicants that Johnston approached agreed toparticipate in her study, 9 did not, and their tapes were erased.

Review procedures differ in detail at different universities, and weencourage researchers who are unfamiliar with their review process to contacttheir campus offices.

COLLECTING DATA AT THE INSTITUTION

In this section we consider some of the steps involved in actual data collection,once all necessary permissions have been secured.


Guideline 16. Take advantage of institutional practices.

Many institutions have practices that make it easier for researchers tocollect data. For example, writing centers often keep copies of the essays thatwere discussed (Thonus, 1999; Williams, 2002). In some cases, having one'swork or talk recorded is a condition of a class, program, or employment. Thestudies by Gibbs of hotel employees (this volume, chapter 7), Davies and Tylerof international teaching assistants (this volume, chapter 5; Tyler & Davies,1990), and Johnston of immigration interviews (2003a, 2003b, in press,discussed earlier) took advantage of situations of this type. In Gibb's study theshadowing of housekeeping employees and the collection of natural data waspart of her job as a materials developer and teacher for the hotel, and it was partof the employees' jobs to participate. In fact, the data collection for the job wentbeyond what Gibbs used for research and included photos, video-taping, audio-taping, copying written materials that employees both received and produced.Similarly, Davies and Tyler report that the requirements of the university weresuch that ITAs [international teaching assistants] and students were required toparticipate in the videotaping for instructional purposes. Tyler writes,"Gathering the data was part of the ongoing activities of the program. Thesubject was enrolled in an IT A class in which his classroom teaching wasvideotaped and analyzed on a regular basis."

Once researchers move from using tape-recorded interactions forinstructional purposes to research purposes, researchers are required to have theparticipants sign release forms. Davies and Tyler report that "the researcherdirectly involved in the videotaping asked the participants for permission to usethe transcript of the videotape for research purposes. Both participants readilyagreed. (In fact they also agreed to let us show the videotape to academicaudiences)."

Guideline 17a. Be prepared for some restrictions.Guideline 17b. Be flexible.

In granting permission for the studies to be carried out, the institutionssometimes place restrictions or requirements on the researchers. Theserequirements are often those of the human subject approvals process. Commonrestrictions included the guarantee of anonymity of the participants and notallowing unauthorized persons access to the data. However others may beparticular to the individual institution such as the restriction of the researchsolely to the portion of the institution under study by the project and not to other


departments or locations where non-participating and/or non-institutionalmembers might appear. An example of this is Gibbs's work done in a hotel,where the participants in the study were hotel employees, but where there was astipulation that no data could be collected where hotel guests might be involved,and permission for taping outside of the housekeeping department had to besecured.

Kerekes and the supervisors at FastEmp negotiated changes at the time thatthe study was proposed. Kerekes reports, "I had a meeting with two of theFastEmp staff and presented them with a preliminary proposal. At our meetingwe negotiated possible changes, after which I submitted a revised writtenproposal to them and their supervisors."

Whereas some institutions restricted access to clients, others restricted thereporting of certain phases of the institutional event. Kerekes's work involvedsome aspects which might be viewed as having legal sensitivity, and so she wasasked not record those portions of the interview, as she reports: "I was notallowed to use as data one portion of the job interviews—the part during whichlegal papers were signed and/or discussed (regarding legal residence,citizenship, and taxes)."

Negotiation and flexibility are key in this type of research. Winn feltconducting post-interviews at the INS office where her research was located wasa potential stumbling block to approval. She told us, "I had originally wanted todo post interviews with immigrant applicants but realized that was a sticky issueso backed down on that." This flexibility allowed her to go ahead with the largerproject. There may be no absolutes as to what is allowed; recall that at anotherINS office, Johnston was permitted to conduct such interviews.

Although we already mentioned Cameron's use of audio-recorders as anexample of unobtrusiveness in the medical setting, this is also an example of arestriction that is institutionally-based. For reasons of client confidentially, novideotaping was permitted.

Guideline 18. Use common sense about what to tape.

We have just discussed aspects of interviews that might be omitted from thepublished data. In addition, in those institutions where personally, politically, orlegally sensitive interactions are undertaken, the researcher might have to asksomeone else to collect the data or might need to forego the data entirely. Thismay be particularly true of medical institutions where, for reasons of privacy ormodesty, a researcher of a different gender might have to gather data, or inpotential health-threatening situations, such as possible contagion in eitherdirection, in which case some other person actually gathers the data. In thecourse of his work at hospitals and clinics, Cameron encountered both of these


situations. In these cases, where possible, trained assistants who were alreadymembers of the institution gathered the data.

As Cameron's work demonstrates, there are situations from which aresearcher may wish to exclude him- or herself. Cameron enlisted the assistanceof a trained female nurse to work at the OB/GYN clinic (because of gender) andat the neurotrauma unit (because of her medical training). Cameron writes:"Also, being a man, I excluded myself from fieldwork and observations in thegynecological clinic and from observations with teenage girls in the HMO clinicwhen it became apparent to the nurse that they were going to have to discussissues of sexuality. ... I also tried to stay out of consultations with individualswho the nurses felt might be highly contagious with something like the measles,even though I had had the measles as a kid. I recall one case where thisoccurred, but it occurred after I was already in the consulting room and hadshaken hands with the patient. So, they let me stay. I also did not do work in theNeurotrauma unit as the setting seemed highly technical and precarious. Hence,the nurse who I trained did the observations there. It turns out that I probablycould have done these observations."

Although often a condition of IRB approvals, offering institutional clientsthe opportunity to request that the recording of the event be terminated is also acommon sense approach to research whether it is required or not. In our researchon advising sessions, we found that students occasionally requested that theportion of the interview following the official advising business not be taped.Once in a while when students were having real difficulty in school, theyrequested that a session not be taped, even though they had signed up for thelongitudinal study. Accommodating participants ultimately benefits the research.

Guideline 19. Know the environment and use appropriateequipment.

Scope out your setting before recording data. Michael Clyne (1994) reportsthat locating areas that were quiet enough to tape in was essential in the factoriesthat he observed. Williams reports moving the writing tutorial upstairs for betterrecording: "We decided that videotaping in the WC itself would be toodisruptive to other writing tutorials (it's a very open space), so we did thesessions in a room upstairs from the center." An open space can also yield poortape quality, with significant background noise.

In the case of the advising sessions, we experimented with our equipment.We had purchased two excellent high quality dual track tape recorders with lapelmicrophones. We had asked advisors to take a tape recorder from the officewhen they picked up the student files for advising. We quickly learned that theadvisors were not comfortable with these large and obviously expensive taperecorders, and that, furthermore, no one really wanted to wear a lapel

216 Bardovi-Hartig and Hartford

microphone, no matter how small. We replaced these tape recorders withinexpensive but good quality portable tape recorders (about 5 x 4 x 1 1/2") withomni-directional microphones. The advisors were much happier with the smallerunits and the data were sufficiently clear without the somewhat intrusivemicrophones (which, although small, were attached to the recorder with a wire,and probably reminded participants that they were being recorded).

As we have already noted, in many cases, the equipment is set up in theoffices in which the interactions were scheduled to take place prior to the event.Not only is this less obtrusive as we discussed earlier, it also assures that theequipment is ready to go when the participants are ready to begin. As Kerekessays, "I had recording equipment set up in the room in which interviews tookplace, and I explained this to the participants before they decided whether or notthey wanted to participate in the study." Johnston reports that after the officeragreed to participate, "I set up his office for videotaping interviews. Onehandheld Sony DCR-PC100 mini digital video camcorder was placed on thewindowsill to capture a side-view of the desk across which the officer and theapplicant spoke during the interview. On top of the desk was a handheldaudiocassette recorder."

Other times portable equipment, such as Yates used, is the most desirable.She bought portable mini recorders and borrowed a video camera from heruniversity.

Guideline 20. Listen to your tapes immediately.

This advice can be found in any field manual, but we include it here so itdoesn't go unsaid: Check your tapes immediately. This stems from at least twoconcerns: equipment failure and interpretation. There could be a problem withthe recording equipment or the setting that you may not have recognized.Checking your tapes will prevent the loss of valuable data. A very commonproblem is old batteries. Put fresh batteries in the tape recorder every day orevery session. Recharge rechargeable batteries and have a backup charged andready to go. Batteries are not the only problem, however. Equipment needs to bemonitored frequently. One of our transcribers pointed out to us that when thevoice activation was set to "on," we were losing important information. Not onlywere the beginnings of turns being cut off, but we were losing the informationon pauses. We made sure that the voice activation was always turned off oncethe problem was identified.

Clyne emphasizes the importance of listening to the tapes immediately forreasons of interpretation, "We found it useful to listen to the tapes immediatelyso that any questions of clarification, identification or interpretation could beattended to as soon as possible."


Johnston pointed out another reason to listen to the tapes immediately: incases where the participants are doing their own taping, to make sure that theparticipants' and the researchers' concepts of the event boundaries are the same.As we reported earlier, the INS officer recorded his own interviews. WhenJohnston reviewed the tapes from the first day, she realized that he had turnedoff the camera when he judged that the official interview was over, which wasafter he received a negative answer to "Any more questions?" As Johnstonlearned, this was sometimes a few minutes before the applicants left the office.She reports, "He was capturing the end of the interview in his judgment, but notthe end of the interaction, which was what I wanted. Reviewing the tapesallowed me to realize that I had not been specific enough in my instructions tohim." Fortunately, the researcher and the interviewer were able to correct thisthe next day, and no further data were lost.

Johnston's story reminded us of our own experience with one of theadvisors who took part in the advising session study. Like Johnston'sinterviewer, he too turned the tape recorder off at the conclusion of the officialportion of the advising session, after the student had received the signedregistration form (Hartford & Bardovi-Harlig, 1992). Like Johnston, we hadhoped to study the closing of the session and any co-membership talk that mightoccur, but unlike Johnston we did not review our tapes immediately, and thuswere unable to make adjustments with that advisor.

Guideline 21. Thank the institution.

When you have completed your study, be sure to thank the institution for itsparticipation. As Dornyei (2003) observes, thanking participants is a basiccourtesy which is often overlooked. The institutions do us a favor to allow us toobserve their daily practices, and they may be more inclined to accommodateother researchers if the basic social courtesies are observed. In addition towriting thank you letters to the participants, Johnston suggested including theparticipants' supervisors as well. In writing to the district supervisor Johnstonthanked him for allowing her to carry out her project and also let him know howhelpful his staff was. She wrote, "I felt that it was good to let him, as asupervisor, know how capably they managed their participation in the projectwhile at the same time performing their regular work duties. I hoped it wouldshow him how research participation can be smoothly integrated with workroutines, and also how professionally his staff treated me." This last point,pointing out that research can be smoothly integrated into work routines, like theextra consideration of writing a thank you letter, helps open the channels ofcooperation with institutions for our own future research as well as that of otherscholars.


AFTER THE STUDY

Once the study has been completed, we approach the issue of dissemination. Wesee at least three types of dissemination. The first is the standard academicpublication, conference presentation, or report, and we will not go into thosehere. The next two might be particular to the setting: sharing information withthe institution and developing practical applications.

Guideline 22. Offer to share information with the institution.

As Sarangi and Roberts (1999, pp. 41-42) point out, there is concern amongresearchers about whether or not institutional research should have tangiblebenefits for the people being studied. Sometimes sharing information is used asa selling point. Sometimes institutions are interested in the information andsometimes they are not. This takes various forms and has different effects.Tyler, Winn, and Johnston all met to share information with officials from thevarious institutions in which their research was located. However, not everystudy results in a sharing of the data with the institution.

In some cases, sharing information with the institution is done throughreports. Winn wrote reports for all parties involved, including INS officers andCitizenship teachers and tutors. Tyler and Davies were part of the institution thatwas being studied. Tyler writes: "I did show portions of the tape to the PhysicsLab supervisor (a faculty member) and the dean of the graduate school. Theytook it as strong indication of the need to continue to support the [ITA]program." Johnston's plan for sharing the data with the institution was twofold,including an informal discussion immediately following the data collectionfollowed by materials development at a later stage. She writes, "My last dayincluded an informal discussion of my impressions and suggestions for changewith my point of contact. After completing data collection, my goal was toanalyze the data for future presentation to academic audiences and to completemy dissertation. I planned to reinterpret my findings with the goal of creatingtraining materials for INS officers and presenting to the INS following thecompletion of my dissertation."

Benefits and Practical Applications

Delivering or developing a practical application for the participating institutionis a specialized form of sharing the information. As we pointed out earlier, onepart of pitching one's project to an institution may be to point out the potentialbenefits of the intended research to the institution itself. This seems to beparticularly true when the researcher is part of the institution.


Although nonnative speakers are often the clients in much institutionalresearch (in this volume see Kerekes, Williams; Bardovi-Harlig & Hartford,1990, 1993, 1996; Hartford & Bardovi-Harlig, 1992; Tarone & Kuehn, 2000;Thonus, 1999; Winn & Seig, 2003), they are increasingly also institutionalrepresentatives in studies of institutional talk and employees in work place talk(in this volume see the chapters by Davies & Tyler, Gibbs, Tarone, and Yates; inother publications, see Cameron & Williams, 1997; Clyne, 1994; Tyler, 1995;Tyler & Davies, 1990), and institutions may be particularly interested inassisting their employees with aspects of language use required by their jobs.

Most of the researchers that we contacted saw a practical application fortheir research, particularly when they teach in the same area as they conductresearch. Winn reports that her work with Seig informs the teaching of U.S.Citizenship (Seig and Winn, 2003). Yates's study of native and nonnativeEnglish speaking teachers in Australian secondary schools informs the advicethat she gives teachers.

The data collected by Tyler and Davies at the physics lab and by Gibbs atthe hotel directly informed teaching and teaching materials. This may be due tothe fact that they were collected in an educational setting. Tyler reportedregularly using the tape itself in subsequent ITA classes. She writes, "I believethe program continues to use it. I also developed a series of strategic role playsbased on the interaction" (See Tyler, 1994.)

Gibbs collected data for the expressed purpose of solving an institutionalproblem (the failure of the novice call-ins) and correcting it through instruction,as part of her job as an ESL teacher at the hotel. Thus the benefits to theinstitution were prearranged as a requirement of her employment. Gibbs usedthe call-ins of the expert English speaking supervisor as an authentic pedago-gical model for novice housekeeping employees. Call-ins made by novices wereplayed in class to help novices identify differences in their call-ins and theywere also used to develop worksheets. As Gibbs reports, "The worksheetsbecame part of the curriculum which was passed on to other teachers andpresented at in-house teacher training sessions to other ESL teachers in theproject. (There were 27 teachers at 43 sites around the Twin Cities involved insimilar teaching scenarios.)"

When asked to reflect on the potential applications of her study ofplacement interviews at the employment agency, Kerekes wrote "the findingsare applicable both to the FastEmp (and other employment agency) staff, whocan learn from my analyses how preconceived notions of communicative stylescontribute to the outcomes of these gatekeeping encounters. The findings canalso be used in ESL and other communication and/or job training courses."

Not all offers of practical application are accepted. Johnston notes thatpractical applications of studying the INS green card interview includeapplications "both for the officers who perform interviews (improving officertraining) as well as the applicants who are interviewed (information on howgreen card interviews are conducted, tips on communicating clearly and helpingthe officer do their job)." Johnston also reports that when her offers of assistance


were not ultimately accepted, she was disappointed, "Interestingly, once Ireceived permission (it took over six months and was very touch and go), therewere no restrictions placed on me—and no requirement to fulfill the offers ofassistance I had made. Once I got through the gate, I was free, both toinvestigate as well as to leave without follow-up. The latter was disappointing; Ihad hoped that once I was finished that there would be interest in what I hadfound or requests for formal follow-up rather than the off-the-cuff discussionsthat concluded my time on site. However, I realize that the INS has enough onits plate without attending to the suggestions of a graduate student. (Note: mydata collection was completed in May 2001, before the repercussions ofSeptember 11, 2001 or the massive institutional reorganization of March 2003,when the division was moved to the Department of Homeland Security.)"

Direct assistance to a particular institution is not the only benefit that can berealized from an institutional study, and thus not a researcher's only chance toput research to practical application if desired. Another type of practicalapplication are published materials or reports such as that written by Cameron(1998), "A language-focused needs analysis for ESL-speaking nursing studentsin class and clinic," and the public presentation by Seig and Winn (2003,March), "Guardians of America's Gate: Discourse-Based Training Lessonsfrom INS Interviews." Such reports reach a much larger audience than theindividual participating institutions. Dissemination of findings and theirpractical interpretation assists practitioners in helping both institutionalrepresentatives and employees and, in turn, their institutional clients. Whetherthe researchers propose the application themselves (Cameron, 1998; Seig &Winn, 2003) or whether they make their results available for others to draw on,dissemination is an important step in the research process.

REFERENCES

Bardovi-Harlig, K., & Hartford, B. S. (1990). Congruence in native and normative conversations:Status balance in the academic advising session. Language Learning, 40, 467-501.

Bardovi-Harlig, K., & Hartford, B. S. (1993). Learning the rules of academic talk: A longitudinalstudy of pragmatic development. Studies in Second Language Acquisition, 15, 279-304.


Bardovi-Harlig, K., & Hartford, B. S. (1997, April). Academic advising sessions with students in anintensive English program. Paper presented at the Eleventh International Conference onPragmatics and Language Learning; Urbana, IL.

Cameron, R. (1998). A language-focused needs analysis for ESL-speaking nursing students in classand clinic. Foreign Language Annals, 31, 202-218.

Cameron, R., & Williams, J. (1997). Sentence to ten cents: A case study of relevance andcommunicative success in nonnative-native speaker interaction in a medical setting. AppliedLinguistics, 18, 415-445.

dyne, M. (1994). Inter-cultural communication at work: Cultural values in discourse. Cambridge,England: Cambridge University Press.

Dornyei, Z. (2003). Questionnaires in second language research: Construction, administration, andprocessing. Mahwah, NJ: Erlbaum.


Douglas, D., & Selinker, L. (1994). Research methodology in context-based second-languageresearch. In E. E. Tarone, S. M. Gass, & A. D. Cohen (Eds.), Research Methodology inSecond-Language Acquisition (pp. 119-131). Hillsdale, NJ: Lawrence Erlbaum Associates.

Hartford, B. S., & Bardovi-Harlig, K. (1992). Closing the conversation: Evidence from theacademic advising session. Discourse Processes, 15,93-116.

Hartford, B. S., & Bardovi-Harlig, K. (1996b, August). Cross-sectional study of the acquisition ofspeech acts by non-native learners of English: The use of natural data in an intensive Englishprogram. Paper presented at the 11th World Congress of the International Association ofApplied Linguistics (AILA), Jyvaskyla, Finland.

He, A. W. (1994). Withholding academic advice: Institutional context and discourse practice.Discourse Processes, 18, 297-316.

Johnston, A. (2003a) A mediated discourse analysis of immigration gatekeeping interviews.Unpublished doctoral dissertation, Georgetown University. Georgetown, Washington, DC.

Johnston, A. (2003b, March). Comembership in immigration gatekeeping interviews: Construction,ratification and refutation. Paper presented at the annual conference of the AmericanAssociation of Applied Linguistics, Arlington, Virginia.

Johnston, A. (in press). Files, forms and fonts: Mediational means and identity negotiation inimmigration interviews. In R. Scollon & P. LeVine (Eds.), Proceedings of the GeorgetownUniversity Round Table. Washington, DC: Georgetown University.

Kerekes, J. (2000, April). The co-construction of a successful gatekeeping encounter: Strategies oflinguistically diverse speakers. Paper presented at the annual conference of Pragmatics andLanguage Learning, Urbana, Illinois.

Kerekes, J. (2001) The co-construction of a successful gatekeeping encounter: Strategies oflinguistically diverse speakers. Unpublished doctoral thesis, Stanford University, Stanford,California.

Kerekes, J. (2003). Distrust: A determining factor in the outcomes of gatekeeping encounters. In S.Ross (Ed.) Misunderstanding in Spoken Discourse. London: Longman/Pearson.

Sarangi, S., & Roberts, C. (Eds.). (1999). Talk, work and institutional order: Discourse in medical,mediation and management settings. Berlin: de Gruyter.

Seig, M. T., & Winn, M. (2003, March). Guardians of America's Gate: Discourse-Based TrainingLessons from INS Interviews. Paper presented at the colloquium on High-stakes GatekeepingEncounters at the American Association of Applied Linguistics, Arlington, Virginia.

Tarone, E., & Kuehn, K. (2000). Negotiating the social service oral intake interview:Communicative needs of normative speakers of English. TESOL Quarterly, 34, 99-126.

Thonus, T. (1999). How to communicate politely and be a tutor, too: NS-NNS interaction andwriting center practice. Text, 19, 253-279.

Tyler, A. (1992). Discourse structure and the perception of incoherence in international teachingassistants' spoken discourse. TESOL Quarterly, 26, 713-730.

Tyler, A. (1994). Effective role-play situations and focused feedback: A case for pragmatic analysisin the classroom. In C. Madden & C. Meyers (Eds.), Discourse and performance ofinternational teaching assistants (pp. 116-131). Alexandria, VA: TESOL.

Tyler, A. (1995). The co-construction of cross-cultural miscommunication: Conflicts in perception,negotiation and enactment of participant role and status. Studies in Second LanguageAcquisition, 17, 129-152.

Tyler, A., & Davies, C. (1990). Cross-linguistics communication missteps. Text, 10, 385-411.Williams, J. (2002). Undergraduate second language writers in the writing center. Journal of Basic

Writing, 27,73-91.Winn, M. (2000). Negotiating borders and discourse: A study of interaction in the US naturalization

interview. Unpublished manuscript. Honolulu, HI: University of Hawai'i atManoa.Winn, M. (under review). Collecting target discourse: The case of the U.S. naturalization interview.

Author Index

Abelson,R., 165, 170Agar, M., 13, 24, 27, 32, 40, 41, 61Ajirotutu, C. S., 8, 32, 99, 129Akinnaso, F. N., 8, 32, 99, 129Aleksa, V., 38, 61Amin,N., 69, 91Aston, G., 11, 12, 29, 32, 56, 61Atkinson, J. M., 199

Bardovi-Harlig, K., 1, 5, 7, 10-3, 15,16, 20-27, 30, 32, 34, 37-40, 47,48, 52, 54, 56, 58, 59, 61, 63, 67,70, 91, 92, 133, 136, 152-155,158, 164, 169, 201, 202, 217, 218,220

Bednarowicz, E., 38, 61Beebe,L., 12, 15, 33, 158, 169Bent, B., 46, 64Bereiter, C., 41, 61Bergman, M. L., 10, 21, 33Bernstein, B., 127, 129Bialystock, E., 136, 155Blau, S., 38, 48, 61Blum-Kulka, S., 9, 33, 34, 48, 61, 67,

69, 72, 91, 94-96Bouquet, E., 38, 61Bouton, L. F., 30, 33Brecke, C., 38, 61Broner,M., 158, 159, 169Brooks, J., 48, 62Brooks, L., 60, 63Brown, J. D., 61, 62Brown, L. F., 38, 62Brown, P., 69, 70, 72, 91, 169, 172Bruffee, K., 37, 62Brunak, J., 70, 93Bruner, J. S., 137, 155Burt,M. K., 136, 155Busher, S., 69, 92Button, G., 12, 33

Cameron, R., 13-17, 21, 25, 26, 29, 30,33, 38, 62, 201, 203, 204, 206,207, 210, 212, 214, 215, 219, 220

Canale, M., 136, 153, 155Capossela, T. -L., 48, 62Churchill, E. F., 16, 18, 26-28, 30, 33Clark, L, 48, 62Clyne, M., 15, 17, 19, 23, 29, 33, 201,

203, 204, 212, 215, 216, 219, 220Cohen, A. D., 11, 15, 19, 33Conley, J., 8, 33Connor, U., 136, 155Cook-Gumperz, J., 8, 33Corder, S. P., 136, 155Cortazzi, M., 69, 92Coupland, J., 90, 92Coupland, N., 90, 92, 127, 129Cumming, A., 41,45, 62Cummings, M. C., 12, 15, 33

Dahl, M., 7, 9-11, 13, 14, 26, 34Damon, W., 38, 62Davies, C. E., 2, 4, 15, 16, 19, 30, 36,

133, 137, 153-155, 201, 213, 218,219, 221

Deby, J., 84, 92De Guerrero, M., 60, 64Denzin, N.K., 108, 129Doany, N., 175, 179, 197, 199Donato, R., 60, 62Douglas, D., 11, 13, 19, 33, 35, 170,

203, 220Dornyei, Z., 21, 32, 52, 61, 217, 220Drew, P., 8, 9, 11, 33, 40, 62Drummond, K., 175, 179, 197, 199Dulay, H. C., 136, 155Duranti, A., 135, 155Dwyer, S., 162, 170Dyehouse, J., 54, 62

223


Eades, D., 169, 170Ellis, R., 16, 18, 26, 28, 33, 135, 136,

155Erickson, F., 8, 13, 19, 26, 33, 38, 55,

62, 99, 129, 135, 153, 155Eslamirasekh, Z., 67, 92

Faerch, C., 136, 155Fairchild, M., 181, 199Fairclough, N., 126, 127, 129Fanshel, D., 8, 35Fen, H. L., 67, 75, 87, 92Ferguson, C., 82, 92Fernando, M. M., 162, 169Ferris, D., 48, 63Fiksdal, S., 37, 62Fisher, S., 8, 33, 38, 62Flowerdew, J., 169, 171Foucault, M., 126, 129Fowler, R., 35Fukushima, S., 67, 92

Gadbow, K., 38, 62Garton, M., 21, 33Gass, S. M., 9-11, 15, 34, 39, 52, 62Gee, J. P., 99, 125, 126, 129Geis, M. L., 10, 11, 34Gibbs, T. L., 2, 5, 162, 166, 167, 169,

171, 175, 194, 199, 201, 209, 213,218, 219

Giles, H., 90, 92, 127, 129Gillam, A., 38, 62Gillespie, P., 38, 62Gillette, S., 162, 170Gimenez, J. C, 169, 171Goffman, E., 12, 34, 68, 92, 197, 199Goldberg, J., 46, 62Goodwin, C, 135, 155Green, G. M., 135, 155Grice, H. P., 12, 34, 135, 155Gubrium, J. F., 1ll, 130Gumperz, J. J., 8, 34, 67, 68, 92, 99,

127, 130, 134, 135, 153-156

Ha, C. T., 67, 92Haastrup, K., 136, 156Hall, J. K., 38,48, 61, 137, 156Halleck, G. B., 128, 130Hansen, J. G., 68, 92

Harlow, L. L., 10, 11, 34Harris, J., 37, 38, 63Harris, M., 37, 38, 48, 62Hartford, B. S., 1, 5, 7, 10, 12, 13, 15,

16, 20-24, 26, 27, 30, 32, 34, 37-40,47, 48, 52, 56, 58, 59, 61, 67,70, 91, 92, 133, 152, 154, 155,158, 164, 169, 201, 202, 217, 218,220

He, A. W., 8, 34, 37, 38, 48, 62, 130,202, 221

Heritage, J., 8, 9, 11, 33, 40, 62, 199Hinds, J., 136, 156Hinkel, E., 69, 92Hinnenkamp, V., 126Hofstede, G., 69, 92Holmes, J., 82, 92, 94Holstein, J. A., 111, 130Hopper, R., 175, 179, 197-199Houck,N., 10, 11, 15, 34, 52, 62House, J., 48, 61, 69, 72, 91, 94-96Huang, G., 38, 61Huang, M. C., 67, 69, 87, 92Hudson, T., 61, 62Hymes, D., 12, 24, 34, 135, 156, 158,

160, 169

Icke, V., 162, 170Irvine, J., 12, 34

Jacobson, W., 169, 171Jefferson, G., 177, 199Jeong, H., 10, 35Jin, L., 69, 92Johnson, M., 175, 179, 197, 199Johnston, B., 10, 34, 201, 202, 205-

214,216-219, 221

Kaplan, R., 136, 156Kasper, G., 2, 6, 7, 9-11, 13, 14, 21,

25, 26, 33, 34, 37,48, 61, 63, 69,72, 91, 94-96, 133, 136, 155, 156

Kennedy, B., 38, 63Kerekes, J., 2, 4, 5, 15, 16, 19, 22, 24-

26, 28, 34, 35, 38, 99, 100, 104,119, 128-130, 201, 203, 206, 207,209-211, 214, 216, 218, 219, 221

Kim, J., 67, 92Kinkead, J., 37, 38, 63

Author Index 225

Koike, D. -A., 67, 92Koran, J. J., 137, 153, 155Kramsch, C. J., 133, 137, 154, 156Kress, G., 35Kubota, M., 67, 92Kuehn, K., 15, 22, 25-28, 30, 36, 38,

64, 162, 163, 169, 171, 203, 218,221

Kuha, M., 10, 35Kuiper, K., 84, 92

Labov, W., 8, 35Lantolf, J., 60, 63Lapkin, S., 60, 63Lave, J., 39, 63Law, J., 38, 63Leary, A., 21, 35Lee-Wong, S. M., 67, 70, 75, 87, 92Levinson, S., 9, 35, 69, 70, 72, 91, 169,

172, 177, 178, 198, 199Li, D., 14, 26, 30, 35Lincoln, Y. S., 108, 129Linde, C., 8, 35Liu, J. L., 68, 92

Mackey, A., 39, 62Maguire, M., 166, 169, 172Mahan-Taylor, R., 154, 155Maier, P., 169, 172Makri-Tsilipakou, M., 46, 47, 63Maley, Y., 169Markee, N., 135, 156, 177, 199Meeuwis, M., 99, 126, 130Mehan, H., 8, 35, 39, 63Messerman, L., 8, 33Miller, L., 169, 171Mir, M., 67, 92Mitchell, C. J., 70, 93Moll, L., 60, 63Moran, M. G., 9, 35Moran, M. H., 9, 35Morgan, M. J., 154, 155Moser, J., 38, 63Murata, K., 46, 47, 63Murphy, B., 10, 35Murphy, S., 38, 63Murray, S., 46, 63Myers, G., 162, 169, 172

Nelms, J., 158, 159, 169Neu, J., 9, 10, 34, 35Newkirk, T., 48, 63Nishiyama, T., 169, 172North, S., 41, 63Norton, B., 90, 93

O'Barr, W., 8, 33Ochs, E., 68, 69, 89, 93O'Donnell, K., 8, 35O'Driscoll, J., 72, 93Ohlstain,E., 11, 15, 19, 33Omar, A. S., 13, 35Ono, R., 21, 35Oxford, R., 136, 156

Parks, S., 166, 169, 172Pathey-Chavez, G. G., 48, 63Phelps, E., 38, 62Philips, S., 8, 35Phillipson, R., 136, 156Powers, J., 38, 63Preston, D., 158, 162, 170Psathas, G., 177, 199

Ranney, S., 164, 165, 170, 172Reynolds, D. W., 154, 155Rintell, E. M., 70, 93Ritter, J., 38, 40, 44, 45, 63Roberts, C., 8, 9, 11, 15, 30, 35, 40, 63,

154, 156, 218, 221Rodriguez, S., 21, 35Rose, K. R., 10, 11, 21, 25, 35, 37, 63,

67, 70, 93Ross, S., 10, 34, 128, 130Rounds, P., 49, 63Rudolph, D., 39, 49, 54, 63Rundquist, S., 158, 170

Sacks, H., 13, 35, 177, 188, 199Salsbury, T., 52, 54, 63Sarangi, S., 8, 9, 11, 15, 30, 35, 40, 63,

99, 130, 218, 221Saville-Troike, M., 135, 156Scarcella, R., 70, 93Scardamalia, M., 41, 61Schank, R., 165, 170Schegloff, E. A., 13, 35, 167, 175-179,

185, 186, 188, 192-195, 197-199


Schiffrin, D., 99, 130Schmidt, M., 170, 173Schmidt, R., 10, 25, 34, 35, 90, 93,

133, 153, 156Scollon, S., 69, 93Searle, J. R., 72, 93Seig, M. T., 201, 202, 218-221Selinker, L., 11, 13, 19, 33, 35, 136,

156, 161, 170, 203, 220Shea, D. P., 67, 93, 99, 126, 130Sheffer, H., 9, 33Shimura, A., 10, 35Shultz, J., 8, 13, 19, 26, 33, 38, 55, 62,

99, 129, 135, 153, 155Siegal, M., 69, 93Silva, T., 38, 62Smith, E., 38, 61So, S., 41, 45, 62Sperling, M., 48, 63Stay, B., 38, 62Storch, N., 37, 60, 63Strauss, T., 38, 61Swain, M., 60, 63, 136, 153, 155Swales, J. M., 11, 13, 36, 39, 64, 159-

161, 170, 181, 182, 199

Takahashi, T., 158, 169Tang, C., 69, 93Tannen, D., 8, 36, 38, 46, 47, 63, 125,

127, 130, 153, 156Tarone, E., 2, 15, 22, 25-28, 30, 36,

38, 44, 64, 157-159, 162, 163,169-171, 182, 203, 218, 221

Thonus, T., 15, 16, 18, 19, 22, 24, 25,30, 36-38, 40, 41, 44, 45, 48, 49,55, 59, 64, 213, 218, 221

Tocalli-Beller, A., 60, 63Todd, A., 8, 33Todd, D., 38, 62Trimble, L., 160Trimbur, J., 37, 64Trosborg, A., 67, 93, 95, 96, 136, 156Tyler, A. E., 2, 4, 15, 16, 19, 24, 25,

30, 36, 37, 64, 133, 134, 137, 138,141, 144, 148, 153-156, 201, 202,213, 218, 219,221

Ulichny, P., 48, 64

van Gelderen, A., 41, 64Varonis, E. M., 11, 34, 52, 62Villamil, O., 60, 64

Wakin,M., 175, 179-181, 197, 199Wallat, C., 8, 36, 38, 63Wang,Z., 10, 35Watson-Gegeo, K., 48, 64Weedon, C., 68, 93Wenger, E., 39, 63West, C., 46, 64Whalen, M., 179, 180, 199Widdowson, H. G., 157, 170Wiemann, J. M., 127, 129Williams, J., 2, 3, 13-17, 21, 25, 26,

29, 30, 33, 37, 38, 40, 54, 62, 64,201, 203, 209, 210, 213, 218-221

Willing, K., 170, 173Winn, M., 201, 202, 206, 211, 214,

218-221Woken, M. D., 11, 13, 36, 39, 64Wolfson, N., 11, 13, 15, 36, 58, 64

Yates, L., 2, 3, 5, 16, 18, 23, 36, 67, 75,93, 201, 203, 204, 206, 207, 209,216,218,219

Yli-Jokipii, H., 9, 36Young, R., 128, 130Young, V., 45, 48, 64Yu, M. -C., 67, 69, 75, 87, 93Yuan, Y., 10, 36Yule, G., 157, 170

Zhang, Y., 67, 93Zimmerman, D., 46, 64, 175, 179-181,

197, 199Zuengler, J., 11, 13, 36, 46, 64

Subject Index

Acquisition of language, 7, 20, 25, 26,31, 61, 136, 153, 162, 166, 198

Adjacency pair, 43, 176-178Aggravator, 45, 48, 49, 58Authentic discourse, 1, 2, 6, 7, 9, 10,

12-15,20-22, 31, 70, 143, 155,158-163, 166, 176, 183, 219

Authority, 1, 3, 46-51, 54-60, 75, 78,80-86, 99, 142-145, 148-152

Backchanneling, 54, 55, 172Background, 2, 3, 21, 37, 39, 69, 71,

82, 90, 125, 162, 171educational, 101, 104, 108, 123ethnic, 107, 113, 160knowledge, 14, 135language, 4, 67, 70-80, 84, 85, 89,91, 109, 121, 122, 129

Backstage, 8Bulge theory, 58

Call-in, 5, 166, 167, 171, 175, 176,181-186, 191-197, 219

Chain of command, 204, 205Chit-chat, 125cold call, 203, 205Comembership, 4, 55-58, 70, 80, 99,

117, 125, 217Comity, 29, 56Comparability (of language samples),

1, 7, 9, 11-14, 19, 20, 26, 28, 31,57, 159, 161, 181, 197

Consequential discourse, 1, 2, 11-14,31, 39

Contextcues, 135-137cultural and linguistic, 4, 72, 90,151, 152, 154, 159influence on institutional talk, 1,19-22, 25, 44, 68, 79, 82, 91, 157institutional, 5, 28, 84, 100, 111,133, 145, 153, 155

interactive, 71, 99, 126, 164social, 11, 69, 70, 138, 158, 168speech event, 7, 27

Contrastive rhetoric, 136Controlled tasks, 7, 10, 11, 21, 22, 25,

70Conversation

as data type, 9-13, 22, 28, 32, 38,135, 184closing, 12, 13, 16, 23, 60, 177,179, 186, 188, 192, 217opening, 5, 12, 60, 144, 167, 171,176-180, 193, 197, 198strategy 16, see also pragmaticstrategy

Cue, 134-137, 146, 172Culture, 4, 81, 89, 100, 136, 149

comparisons, 90, 134, 135, 171diversity, 39, 101, 109, 154, 164,173institutional talk and, 28, 42, 48,99, 110, 113, 126, 147, 155, 165L1 influence, 69, 133, 150, 153norm, 4, 21,67, 69, 84, 90, 133,140, 145, 147, 150-152, 158, 161references in talk, 56, 84, 96specific, 14, 26

Demeanor, 68Detection phase, 41Diagnosis phase, 40-44, 47, 57, 59Directive, 3, 16-18, 23, 27, 40-43, 47-

49, 57-60, 67-76, 79-86, 89, 90,97, 164, 172, 186-188, 195, 197

Directness, 1, 3, 72, 82, 86, 113, 134,142, 143, 152, 162, 170, 173

Discourse community, 2, 5, 37, 67, 68,90, 91, 133-137, 140, 146, 157-168, 172, 175, 179, 181, 182, 194,196, 198

Discourse completion tasks (DCTs), 9,10, 70, 133, 159, 165

227


Discourse strategy, see pragmaticstrategy

Discourse style, 1, 2, 46, 73, 77-91, 99,118, 125, 134, 158, 171,219

Disfluency, 179, 181Dispreferred second, 177, 179, 187,

188, 195Dominance, 1-3, 44-48, 51-55, 60, 99,

126Downgrader, 52Downtoning, 85, 87, 96

Elicitation tasks, 1, 2, 7, 10, 21, 22, 70,133, 159

English as a Second Language (ESL),3, 36, 71, 72, 175, 176, 182-184,196, 202, 219, 220

English for Specific Purposes (ESP), 5,157-163, 168-170

Equality, 1, 38, 60Escalation, 1, 20, 23, 49, 137Ethics, 30, 72, 142, 146, 151, 204, 208Ethnicity, 106, 107, 113, 121-127, 160Evaluative exchange, 19Expert

model, 191, 219status, 44, 55, 57, 184, 185, 203vs. novice, 26, 38, 114, 159, 175,176, 183, 192-196

Expertise, 2, 160, 188acquiring, 28, 32, 163asserting, 19, 40, 45, 47, 60, 113,128in discourse communities, 5, 161-168, 172, 181, 182, 205, 207

Feedback, 21Filler, 116Floor management, 3, 54, 55, 60, 115,

146-152, 178, 184, 189-192Foreigner talk, 82Frontstage, 8

Gatekeeping, 99, 112, 125-128, 205,219

Gender, 2, 3, 6, 22, 72-84, 106, 107,121-127,214

Genre, 5, 26, 28, 134, 157-168, 172,181, 182, 194-198

Goaldiscourse, 5, 9, 11, 13, 15, 20, 25,31, 37, 40, 55, 136, 154, 166, 168,180, 197institutional, 110, 157, 160, 161,167, 181, 182, 192, 194, 207language learners', 59, 61, 90, 163long-term, 70, 124research, 2, 208, 218vs. outcome, 1, 24

Identification phase, 41, 177-181, 185,187-190, 193-198

Imposition, 1, 20, 84, 95, 172Inductive/collaborative strategy, 20,

134, 137, 138, 142, 145-148, 152Inference, 135, 157Inflectional morphology, 41Input, 3, 16, 28, 164, 166Institutional support, 209, 218Institutional talk, 1-3, 5-9, 11-16, 19-

22, 24-29, 31, 32, 37-41, 44, 47,59, 133, 152, 153, 198, 201, 202,218defined, 7, 8, 38, 41generalizability, 28-33, 159, 181identity and, 3, 8, 9, 23, 67-70,77-79, 84, 86, 89-91, 125, 181,182, 185, 194membership and, 19, 181, 182replicability, 13, 22, 31, 70rules, 1, 20, 25-28

Integrative motivation, 163Interaction, 24, 69, 99, 136, 158, 163,

170, 172in institutional research, 202, 204,209, 212, 213, 216-219in job interviews, 110-113, 127in phone calls, 176, 179, 188, 197of L2 writers, 38, 40, 57, 58, 61with international TA, 140-143,148-151

Interactional/interactivecompetence, 137, 153context, 37, 68, 70, 71, 126, 164,173language samples, 11, 12, 14, 20,31perspective, 133-135, 154

Subject Index 229

status development, 44, 45, 48, 54style, 60, 125

Intercommunication, 5, 181Interlanguage pragmatics research

defined, 7, 157Interlocking organization, 177, 178Interpreter, 101, 120, 164Interruption, 3, 45-47, 57, 65, 112, 131

Job shadowing, 175, 176, 183, 184,213

Job type, 2, 106, 107, 113, 121-126

L1 pragmatics, 4, 136, 152, 154L2 learner, see language learnerL2 pragmatics, 3, 4, 19, 22, 26, 61,

109, 120-122, 128, 162L2 writer, 2, 3, 37-60Language learner, 22, 35-28, 30, 31,

37-39, 54, 61, 68, 90, 133, 136,154, 159, 163-165, 168, 201

Language needs analysis, 30, 182, 220Language play, 159, 162, 172Language transfer, 134, 142, 145, 151-

154,173Language use, 3, 5, 7, 15, 20, 21, 26,

67-70, 84, 86, 89-91, 109, 121,157, 160-163, 168, 219

Language variety, see genreLeading move, 51-54Lexical choice, 5, 81, 86, 87, 96, 97,

131, 135, 140, 144, 146, 152, 153,163, 181

Limited proficiency English speaking(LPES), 175, 176, 183, 184, 188,189, 192-198

Miscommunication, 3, 16, 19, 22, 25,28, 69, 99, 113, 127, 137, 153,154, 164, 171, 172, 198, 208

Mitigator, 3, 16, 45, 48, 49, 52-55, 58,59, 69-91, 95-97, 143, 167, 172

Modality, 48, 49, 54, 81, 95, 144, 165,172

Modeling, 5, 167, 182-185, 191, 194-197, 219

Motivation, 19, 40, 110, 135, 147, 163Mutuality, 60

Native English (NE), 2, 3, 37-40, 44-61

Negotiation, 4, 11, 12, 19, 21-23, 37,40, 54, 58-61, 67-70, 79, 81, 171,188, 190, 214

Networking, 112Neutral language, 42, 43Nonreciprocal role, 3, 60Norms of interaction, 9, 12, 21, 25, 37,

60, 82, 134, 136, 142-147, 150-152, 158-160, 162, see also normunder culture

Open-ended question, 113, 146Opening move, 5, 175-180, see also

opening under conversationOptimal convergence, 90Organization pattern, 134, 135, 137,

160, 162, 165Outcome, 1, 2, 4-6, 10, 11, 20, 21, 24,

25, 31, 55, 99, 112, 113, 121,202,219

Overlapping speech, 47, 65, 131, 138,187, see also simultaneous speech

Packing, 178Parallelism, 142Politeness marker, 20, 23, 31, 48, 59,

70, 73, 74, 79-83, 88, 90, 96, 172,181, 188, 191, 192

Postsession interview, 10, 19, 20, 28,30, 38, 47, 48, 54, 59, 72, 83, 100,108, 109, 134, 140,211,214

Pragmatic competenceassessment, 109, 120, 121, 129,159, 161, 168, 198factors, 52, 90, 136, 137, 153L2 learner, 2, 60, 61, 75, 78, 84-87, 133, 144, 154, 158, 159, 165native speaker, 67, 157-159

Pragmatic effect, 5, 157Pragmatic feature, 25, 158Pragmatic strategy, 2, 3, 133-137, 140,

142, 145-154, 165, 171, 173, 188Preferred second, 177, 187, 188Pretext, 99, 126Production questionnaire, 1, 10-12, 21,

22

230 Bardovi-Harlig andHartford

Proficiency level, 25, 26, 37, 61, 78,84, 87-89, 120, 121, 128, 153,154, 183

Race, see ethnicityRefusal, 10, 12, 16, 54, 178, 190Rejection, 52-54, 119, 171, 177, 179,

186, 188, 195Report writing, 24, 40, 41Rescue, 47, 57Research design, 1, 2, 152, 153, 158,

159, 168, 169, 201Research guidelines, 6, 161, 201-218Review board process, 211, 212, 215Role play, 9-13, 21, 22, 70, 172, 182,

183, 219Role

in institutional talk, 9, 18, 28-31,44, 59, 75, 99, 165, 185, 188negotiation, 55, 58, 60, 61, 81social, 68, 91, 110, 163speaker, 2, 37-40, 48, 56, 78, 134,136, 150-152, 171, 172

Rules (of discourse), 5, 37, 61, 162,182, 194-198, see also rules underinstitutional talk

Script, 163-165, 168, 171-173Second language learner, see language

learnerSequential contributions, 22, 82Shared knowledge, 13, 67-69, 135,

157, 163, 168Simulated tasks, 11-13, 70Simultaneous speech, 47, 55, see also

overlapping speechSituated language, 2, 7, 31, 135, 136,

140, 146Smalltalk, 17, 55, 56, 60Social constructivism, 68Social distance, 1, 23, 47, 72-74, 78-

91, 94, 95Socioaffective consequence, 12, 13Solidarity, 1, 3, 47, 55, 60, 72, 75, 78-

90Speech act, 3, 10, 13, 15, 17, 18, 22,

23, 27, 31, 45, 55, 60, 133, 136,143, 154, 158, 159, 163-172, 195

Speech act analysis, 14, 22, 133, 173

Speech behavior, 14, 16, 69, 91, 110,121, 153, 158, 159, 165

Speech event, 7, 14, 15, 24, 29, 31, 60,161, 165, 181, 194, 198

Spontaneous conversation, 9-14, 136,175, 183

Status, 1, 6, 12, 14, 22, 28, 29, 3730,44, 47-49, 52-59, 159, 164, 165

Stereotype, 100, 113, 126, 127, 219Subjectivity, 68Subject specialist informant, 19, 203Success, 3, 153, 205, 210

in discourse community, 5, 175in institutional talk, 13, 16, 17, 22,24, 99, 108, 109, 112-114, 118,120-129, 167, 184-198L2 learner, 37, 176, 182native vs. nonnative speaker, 1, 2,4, 14, 25

Teachers of English to speakers ofother languages (TESOL), 12

Third place, 133, 154Timing, 1, 20, 23, 31, 127Topic introduction, 179-181, 184-198Topic shift, 51, 144Trust, 1, 99, 102, 104, 114, 121, 127,

128Turn length, 3, 45, 57, 60, 116Turns at talk, 22, 31, 54, 55, 58, 178-

180, 187-190, 198, 216Turn-taking, 9-12, 17, 27

Upgraders, 3, 49, 59

Wait time, 47Workplace talk, 8, 14, 17, 38, 204, 218,

see also institutional talk

interlanguage pragmatics bardovi-harlig&hartford lb

Documents